<small><font color="red">IMPORTANT:</font> The COLD FUSION (GAIN+Unsloth) method of training maintains 99% of performance of BF16, at both 8 bit and 4 bit levels.
Model exceeds all Qwen 3.8, 3.6 and 3.5 27B critical core benchmarks. A model that gets down to business faster, with less "talking" and is smarter too.</small>
<h2>Qwen3.8-27B-Cold-Fusion-GAIN-V1.1-NM-DAU-NEO-MAX-MTP-GGUF</h2>
<img src="cannonball.webp" style="float:right; padding:10px;">
Cold Fusion has 1/5 (as low as 1/10 in some cases) to 1/2 the thinking tokens (vs reg Qwen 3.8) across all 3 modes of operation, and it is faster and smarter too
created using the COLD FUSION method of training.
This is a high detail focused model, with tuning specific to address over reasoning/over thinking and excessive token consumption.
EXAMPLE generations at the bottom of the page.
A Colab between myself (tuning, COLD Fusion), Nightmedia (benching), and TeichAI (Datasets).
The strict goals of this model creation were:
- Increase the general model intelligence and problem solving abilities.
- Reduce thinking block size -> now at 1/10 to 1/2 the size [median reduction: 2/3 roughly].
- Reformatting the thinking block, as well as improving it.
- Speed up token generation, especially MTP.
- Ensure all updates work with all three modes of thinking.
- ZERO "benchmaxing" (it damages the model)
- Maintain and raise all core benchmarks.
<B>COLD FUSION ("Gain" + "Unsloth") TRAINING: </B>
COLD FUSION (GAIN+UNSLOTH) training tech which was invented by my team during the R & D
of "Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic" (2100+ likes, 3 million + downloads, 60+ quant repos):
https://huggingface.co/DavidAU/Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF
The "GAIN" is the core invented component, then coupled with Unsloth's trainers/systems => AKA -> COLD FUSION.
The "GAIN" method (programming) automatically (and dynamically) changes training on a per sample basis in real time during training AS THE MODEL LEARNS.
The method improved metrics as well as overall model performance without overcooking or damaging the model.
This has also resulted, in the strongest and most stable model at both 4 bit and 8 bit and made 4 bit performance 99% of 8 bit performance too.
<B>TESTING:</B>
Testing and benching was done at each stage to ensure quality.
You can also see benchmarks below too for this model, Qwen 3.6 27B, and Qwen 3.5 27B.
HOWEVER, the final testing was HUMAN testing. A trust, but verify approach.
Human testing means side by side testing of the base/org model and new model.
Features:
- Improved instruction following.
- Overall increase in general intelligence and problem solving.
- Better thinking/reasoning with far smaller thinking/reasoning blocks, output generation will also be compressed by default in many cases.
- Even lower/lowest quants are exceptional.
- No corruption or change to Team Qwen's exceptional model - everything is there.
- Vision
<B>IMPORTANT:</B>
This model, like regular Qwen 3.8 27b, supports THREE modes of reasoning : xhigh (default), medium and low [see info in Qwen 3.8 section below].
Reduction in thinking tokens/reasoning block size extends across all three modes of operation.
Likewise detail levels extend to all three modes too, even with reduced thinking/reasoning block the OUTPUT detail will remain high.
To REDUCE thinking block[s] further, increase the level/detail of your instructions/prompts - it only takes a little bit more here so the model has to guess / reason a little bit less.
Also, generally within the same chat additional reasoning blocks will also be reduced from typical Qwen levels.
Also note that the modification of "reasoning" is a major change to the model please carefully test it for your use case(s).
<B>Modification of REASONING:</b>
If you AI app does not support a "switch" you can manually modify the JINJA template.
The default setting is "xhigh" ; to change to medium or low use:
{%- set reasoning_effort = 'medium' %}
OR
{%- set reasoning_effort = 'low' %}
Place this at the VERY TOP of the jinja template.
In LMStudio you can access this in DEV mode, and switch off the "advanced updates" option.
Other AI apps may vary.
You can also make your own quants from source here:
https://huggingface.co/DavidAU/Qwen3.8-27B-Cold-Fusion-GAIN-V1.1
Just modify the "chat-template.jinja" (in NOTEPAD or similar) AND the token-config.. json file too (or delete the "chat template" from this file).
ADVANCED:
Qwen 3.8 uses System prompt injection control by the Jinja template to control reasoning levels.
If you set it at "medium" this turns off injection [ie: no system prompt is injected]
You can then set a "reasoning" system prompt yourself.
The other option: Modify the jinja itself and the system prompt(s) to better tune reasoning to your use cases.
<B>Regular and MTP GGUFS:</b>
All quants (regular and MTP) are NEO IMATRIX, which improve accuracy of the quants by an additional 2-4% over normal GGUFs as well as long context performance.
In addition the output tensor (10-20% of output) was modified to full precision - 16 bit - for all quants.
"MTP" GGUFS (multi-token prediction):
- "MTP" GGUFS will have "MTP" in the name as a suffix.
- I have also set the MTP tensors to Q8_0 precision for all quants.
- To get better performance keep temp 1 or less (higher temps degrade MTP performance).
- Likewise with rep pen ; keep at 1 (off). If you raise it performance will suffer.
- If you see "token acceptance" rates BELOW 50% (predict 2 tokens) switch to normal quants.
I have also added two "LOW" quants, with "LOW" in the name:
- IQ4_XS and Q6_K
- These are for max speed / reduced VRAM and without MTP/OT mods.
- Performance may be slightly lower than the reg "MAX" quants.
SPEED:
- On Q4_K_S (4bit) quant, regular GGUFs are about 75 t/s, whereas MTP GGUFs (acceptance at 60%, 2 tokens) can exceed 90 T/S. (5090, Windows 11, testing in LMStudio)
- Speeds will vary depending on GPU(s), AI app, O/S (Linux/Mac will generally be faster) and hardware.
- "MTP" quants speeds will vary ; for creative/complex and/or temps over 1 use regular GGUFs for better performance.
I suggest you download at least one of each - regular and MTP gguf(s) - and test them for your use case(s).
If you get "token acceptance" (predict 2 tokens) with MTP quant(s) BELOW 50% (this means regular quants will run faster), then regular GGUF(s) will actually perform better - ie faster.
MTP quant(s) can in some cases run faster as the token window fills up and/or in multi turn chats.
Note there is NO other diffence between the quants type besides speed: both will do the same job.
<B>Model:</b>
- 256k context
- Gguf quants run in all standard AI apps.
- Vision is activated, but you need to download separate "mmproj" file (ONE) to use it.
<B>VISION:</B>
- Vision (images) tested.
- You need an "mmproj" (just one) of these downloaded too, and placed in the same folder as the GGUF for images.
<B>Qwen Model Settings (suggested):</B>
- Thinking mode for general tasks: temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0
- Thinking mode for precise coding tasks (e.g. WebDev): temperature=0.6, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0
- Instruct (or non-thinking) mode: temperature=0.7, top_p=0.80, top_k=20, min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0
- Context window min from 8k to 16k ; suggest 24k to 32k even with reduced reasoning blocks.
<h2>BENCHMARKS by Nightmedia</h2>
Important note on Qwen 27B 3.8 bench VS Qwen 3.6/3.5 27B versions:
Based on my testing / Qwen's own statements, community statements (ie localllama) and extended benchs for 3.8-27B version (team Qwen) this model is more focused on
deeper thinking, coding and agentic functions than previous Qwen versions.
arc/c arc/e boolq hswag obkqa piqa wino
Qwen3.8-27B-Cold-Fusion-GAIN-V1.1 [non heretic]
mxfp8 0.655,0.838,0.898,0.751,0.498,0.807,0.738
mxfp4 0.645,0.833,0.887,0.740,0.496,0.799,0.732
Qwen3.8-27B-Instruct: [base, non heretic]
mxfp8 0.591,0.782,0.896,...
mxfp4 0.581,0.771,0.889,0.738,0.442,0.798,0.713
Qwen3.6-27B-Instruct: [base, non heretic]
mxfp8 0.647,0.803,0.910,0.773,0.450,0.806,0.742
Qwen3.6-35B-A3B-Instruct [base, non heretic]
mxfp8 0.581,0.757,0.892,0.751,0.428,0.803,0.688
Qwen3.5-27B-Instruct: [base, non heretic]
mxfp8 0.557,0.711,0.868,0.533,0.452,0.706,0.695
NOTES:
- Models are tested in "Instruct" mode because this generally works better with the testing harness.
- Testing via "thinking" mode also shows the metrics (and changes) but not the true extent.
- In actual fact when the model IS in thinking mode, it will exceed INSTRUCT benchmark scores in most cases.
- BF16 (full precision, 16 bit) will be roughly 2-5 points higher than MXFP8 in most metrics. Some metrics may be slightly higher than this.
<B>Settings: CHAT / ROLEPLAY and/or SMOOTHER operation of this model:</B>
In "KoboldCpp" or "oobabooga/text-generation-webui" or "Silly Tavern" ;
Set the "Smoothing_factor" to 1.5
: in KoboldCpp -> Settings->Samplers->Advanced-> "Smooth_F"
: in text-generation-webui -> parameters -> lower right.
: In Silly Tavern this is called: "Smoothing"
NOTE: For "text-generation-webui"
-> if using GGUFs you need to use "llama_HF" (which involves downloading some config files from the SOURCE version of this model)
Source versions (and config files) of my models are here:
https://huggingface.co/collections/DavidAU/d-au-source-files-for-gguf-exl2-awq-gptq-hqq-etc-etc-66b55cb8ba25f914cbf210be
OTHER OPTIONS:
-
Increase rep pen to 1.1 to 1.15 (you don't need to do this if you use "smoothing_factor")
-
If the interface/program you are using to run AI MODELS supports "Quadratic Sampling" ("smoothing") just make the adjustment as noted.
<B>Highest Quality Settings / Optimal Operation Guide / Parameters and Samplers</B>
This a "Class 1" model:
For all settings used for this model (including specifics for its "class"), including example generation(s) and for advanced settings guide (which many times addresses any model issue(s)), including methods to improve model performance for all use case(s) as well as chat, roleplay and other use case(s) please see:
[ https://huggingface.co/DavidAU/Maximizing-Model-Performance-All-Quants-Types-And-Full-Precision-by-Samplers_Parameters ]
You can see all parameters used for generation, in addition to advanced parameters and samplers to get the most out of this model here:
[ https://huggingface.co/DavidAU/Maximizing-Model-Performance-All-Quants-Types-And-Full-Precision-by-Samplers_Parameters ]
Qwen3.8-27B
[!Note]
This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format.
These artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, TokenSpeed, etc.
[!Tip]
For users seeking managed, scalable inference without infrastructure maintenance, the official Qwen API service is provided by Qwen Cloud.
In particular, Qwen3.8-27B will be available as a hosted version with more production features, e.g., 1M context length by default, official built-in tools. For more information, please refer to the Qwen3.8-27B Overview. The service is coming soon. Stay tuned for updates.
Following the widespread community adoption of the Qwen3.5 and Qwen3.6 series, we are pleased to introduce Qwen3.8, the most capable generation in the Qwen open-model family to date.
Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability.
Qwen3.8 Highlights
Qwen3.8-27B features the following enhancements:
- Core Capabilities: Comprehensive improvements across coding, professional work, research, and long-horizon agentic tasks.
- Agent Execution: Stronger autonomous planning and better handling of environment feedback, leading to more reliable end-to-end task completion.
- Downstream Compatibility: Broader support for popular harnesses and development tools, making it easier to integrate into your existing stack.
- Flexible Thinking Control: Thinking mode is on by default and can be disabled per request; reasoning depth can be tuned with
reasoning_effort, and reasoning context from historical messages is retained via preserve_thinking.
- Vision-Language Understanding: Native support for image and video understanding, from STEM diagrams and documents to hour-scale videos.
Model Overview
- Type: Causal Language Model with Vision Encoder
- Training Stage: Pre-training & Post-training
- Language Model
- Number of Parameters: 27B
- Hidden Dimension: 5120
- Token Embedding: 248,320 (Padded)
- Number of Layers: 64
- Hidden Layout: 16 × (3 × (Gated DeltaNet → FFN) → 1 × (Gated Attention → FFN))
- Gated DeltaNet:
- Number of Linear Attention Heads: 48 for V and 16 for QK
- Head Dimension: 128
- Gated Attention:
- Number of Attention Heads: 24 for Q and 4 for KV
- Head Dimension: 256
- Rotary Position Embedding Dimension: 64
- Feed Forward Network:
- Intermediate Dimension: 17,408
- LM Output: 248,320 (Padded)
- MTP (Multi-Token Prediction): trained with multiple steps
- Context Length: 262,144 natively and extensible up to 1,000,000 tokens.
Benchmark Results
Text Performance
<style>
.vl-table th{font-size:15px!important;line-height:1.2}
.vl-table td:not(.benchmark-cell):not([colspan]){font-size:15px;line-height:1.2;vertical-align:middle}
.vl-table .benchmark-cell{padding:12px 10px 12px 18px!important;vertical-align:middle}
.vl-table .benchmark-capability{font-size:15px;font-weight:600;line-height:1.22;color:#171717}
.vl-table .benchmark-name{margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B}
.vl-table .metric-stack{display:flex;flex-direction:column;gap:7px;padding:3px 0}
.vl-table .metric-label{font-size:10px;font-weight:400;line-height:1.1;color:#777}
.vl-table .metric-value{margin-top:2px;font-size:15px;line-height:1.15;color:#171717}
</style>
<div style="font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Roboto,sans-serif;max-width:1200px;margin:0 auto;padding:16px 0">
<table class="vl-table" style="width:100%;table-layout:fixed;border-collapse:collapse;font-size:13px">
<thead><tr>
<th style="padding:10px 7px;text-align:left;font-weight:600;border-bottom:2px solid #0A2EFE;color:#0A2EFE"></th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;background:rgba(10, 46, 254, 0.08);">Qwen3.8-27B</th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;">Qwen3.6-27B</th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;">Qwen3.7-Plus</th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;">Muse Glimmer-30B</th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;">Opus4.6 Max</th></tr></thead>
<tbody>
<tr><td colspan="6" style="padding:8px 12px;font-weight:600;color:#0A2EFE;border-bottom:1px solid rgba(10, 46, 254, 0.2);background:#D6DAFC">Coding</td></tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Agentic terminal coding</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">Terminal Bench 2.1 (Terminus)</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;">73.0</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">63.4</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">64.0</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">51.7</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>78.2</strong></td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Agentic coding</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">SWE-bench Pro</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>61.7</strong></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">53.5</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">57.6</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">51.2</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">53.4</td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Repo-level code generation</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">NL2Repo-Bench</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;">42.3</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">36.2</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">41.1</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>47.6</strong></td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Agentic coding</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">DeepSWE 1.1</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>42.2</strong></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">13.3</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">14.2</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Software engineering</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">QwenSWEBench</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>79.0</strong></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">49.3</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">59.2</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">63.8</td>
</tr>
<tr><td colspan="6" style="padding:8px 12px;font-weight:600;color:#0A2EFE;border-bottom:1px solid rgba(10, 46, 254, 0.2);background:#D6DAFC">Agent</td></tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Long-horizon office work</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">CoWorkBench</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>70.7</strong></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">61.0</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">65.1</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">68.2</td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Professional job tasks</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">JobBench</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>33.4</strong></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">21.8</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">27.6</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Frontier agentic tasks</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">Agents' Last Exam</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Pass@1</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>20.4</strong></div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Score</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>42.9</strong></div></div></div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Pass@1</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">10.6</div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Score</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">27.3</div></div></div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Pass@1</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">13.2</div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Score</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">33.6</div></div></div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
</tr>
<tr><td colspan="6" style="padding:8px 12px;font-weight:600;color:#0A2EFE;border-bottom:1px solid rgba(10, 46, 254, 0.2);background:#D6DAFC">General</td></tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Instruction following</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">IFBench</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>79.5</strong></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">69.1</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">79.1</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">77.0</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">62.5</td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Scientific reasoning</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">GPQA Diamond</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;">89.2</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">87.8</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">90.3</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">83.5</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>91.3</strong></td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Multidisciplinary reasoning</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">HLE</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;">30.8</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">24.0</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">34.7</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">22.0</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>40.0</strong></td>
</tr>
<tr>
<td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Competitive coding</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">LiveCodeBench v6</div></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>90.3</strong></td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">83.9</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">89.6</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td>
<td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">88.8</td>
</tr>
</tbody>
</table>
<div style="margin-top:12px;font-size:11px;line-height:1.5;color:rgba(0,0,0,0.72)">
<ol style="margin:0;padding-left:20px">
<li>SWE-bench Pro: Except for Opus4.6 Max, which uses the officially reported score, all models are evaluated with the Claude Code harness at temp=1.0, top_p=0.95, and a 256K context window. Problematic tasks were corrected, and all baseline models were re-evaluated on the refined benchmark.</li>
<li>NL2Repo-Bench: Evaluated with the Claude Code harness. To prevent reward hacking, we disable Bash commands that attempt to access the specific repository, such as pip download, pip install, and git clone.</li>
<li>DeepSWE 1.1: Evaluated with the Claude Code harness at temp=1.0, top_p=0.95, and a 256K context window.</li>
<li>QwenSWEBench: In-house coding benchmark for evaluating models' software engineering capabilities. Evaluated with the Claude Code harness. Reporting avg@3 with an 8-hour timeout, max_tokens=32,768, temperature=1.0, and a 256K context window.</li>
<li>CoWorkBench: In-house cowork benchmark for evaluating long-horizon tasks across computer science, finance, law, medical, and other productivity domains.</li>
<li>HLE: Judged by GPT-4o.</li>
<li>The best result in each row is shown in bold.</li>
<li>Empty cells (--) indicate that results are not yet available or not applicable.</li>
</ol>
</div>
</div>
VL Performance
<div style="font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Roboto,sans-serif;max-width:1200px;margin:0 auto;padding:16px 0">
<table class="vl-table" style="width:100%;table-layout:fixed;border-collapse:collapse;font-size:13px">
<thead><tr><th style="padding:10px 7px;text-align:left;font-weight:600;border-bottom:2px solid #0A2EFE;color:#0A2EFE"></th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;background:rgba(10, 46, 254, 0.08);">Qwen3.8-27B</th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;">Qwen3.6-27B</th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;">Qwen3.7-Plus</th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;">Muse Glimmer-30B</th><th style="padding:10px 7px;text-align:center;font-weight:500;border-bottom:2px solid #0A2EFE;color:#0A2EFE;font-size: 14px;width:14.00%;">Opus4.6 Max</th></tr></thead>
<tbody>
<tr><td colspan="6" style="padding:8px 12px;font-weight:600;color:#0A2EFE;border-bottom:1px solid rgba(10, 46, 254, 0.2);background:#D6DAFC">Agentic Multimodal Intelligence</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Computer use</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">OSWorld-Verified</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>84.3</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">63.9</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">73.3</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">65.9</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">72.7</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Browser use</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">WebArena-Verified</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>64.8</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">48.8</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">55.3</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Mobile use</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">AndroidWorld</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>81.9</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">70.3</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">81.0</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">62.0</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Application recreation</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">RecreationBench</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>47.1</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">29.8</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">30.2</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Multimodal tool use</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">ClawEval-MM</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Pass@3</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>57.4</strong></div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Average</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">56.9</div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Pass@3</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">42.6</div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Average</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">50.4</div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Pass@3</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>57.4</strong></div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Average</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>60.1</strong></div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Pass@3</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">52.5</div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Average</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">54.7</div></div></div></td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Multimodal software engineering</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">SWE-MM</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>38.6</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">25.7</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">30.0</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">27.1</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Visual web development</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">Vision2Web</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>62.9</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">45.0</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">42.1</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td></tr>
<tr><td colspan="6" style="padding:8px 12px;font-weight:600;color:#0A2EFE;border-bottom:1px solid rgba(10, 46, 254, 0.2);background:#D6DAFC">General Multimodal Intelligence</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Visual math problem solving</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">MathVision</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">90.0</div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">With CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>94.6</strong></div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">85.1</div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>90.3</strong></div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">65.5</div></div></div></td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">General visual reasoning</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">BabyVision</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>65.7</strong></div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">With CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>85.6</strong></div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">28.9</div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">64.7</div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">With CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">70.4</div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">12.6</div></div></div></td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Scientific chart analysis</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">CharXiv (RQ)</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">83.7</div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">With CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>90.2</strong></div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">78.4</div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717"><strong>85.8</strong></div></div><div style="margin-top:7px"><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">With CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">85.9</div></div></div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">78.8</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><div class="metric-stack" style="padding:3px 0"><div><div class="metric-label" style="font-size:10px;font-weight:400;line-height:1.1;color:#777">Without CI</div><div class="metric-value" style="margin-top:2px;font-size:15px;line-height:1.15;color:#171717">66.0</div></div></div></td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Document intelligence</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">OmniDocBench 1.5</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;">91.1</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">89.4</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>91.4</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">75.8</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">86.6</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Real-world perception</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">RealWorldQA</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;">85.9</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">84.1</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>86.9</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">73.9</td></tr>
<tr><td class="benchmark-cell" style="padding:7px 7px;padding-left:20px;border-bottom:1px solid rgba(128, 128, 128, 0.15);"><div class="benchmark-capability" style="font-size:15px;font-weight:600;line-height:1.22;color:#171717">Embodied intelligence</div><div class="benchmark-name" style="margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;color:#6B6B6B">ERQA</div></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 46, 254, 0.08);vertical-align:middle;font-size:15px;line-height:1.2;">65.5</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">62.5</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;"><strong>69.8</strong></td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">--</td><td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);vertical-align:middle;font-size:15px;line-height:1.2;">40.8</td></tr>
</tbody>
</table>
<div style="margin-top:12px;font-size:11px;line-height:1.5;color:rgba(0,0,0,0.72)">
<ol style="margin:0;padding-left:20px">
<li>MathVision, BabyVision, and CharXiv (RQ): Where both settings are available, cells report “Without CI” and “With CI” separately; otherwise, only the available setting is shown. A small number of incorrect ground-truth annotations in MathVision and CharXiv (RQ) were corrected following manual verification, and all reported scores on those benchmarks were computed using the corrected annotations.</li>
<li>MathVision: Qwen3.8-27B is evaluated using the fixed prompt: “Please reason step by step, and put your final answer within <code>\boxed{}</code>.” For the remaining models, we report the higher score from two prompt variants—one with and one without the <code>\boxed{}</code> formatting requirement.</li>
<li>WebArena-Verified: Scores are computed with the official WebArena-Verified grader under the OSWorld scaffold.</li>
<li>RecreationBench: An in-house, long-horizon application-recreation benchmark designed to evaluate hybrid-agent capabilities across five platforms: desktop (Ubuntu, macOS, and Windows), mobile (Android), and the web.</li>
<li>ClawEval-MM: Scores are reported as “Pass@3 / average score.” Pass@3 is the percentage of tasks passed in at least one of three trials; the average score is the mean benchmark score across the three trials.</li>
<li>Vision2Web: Scores are averaged across the frontend, webpage, and website categories. Evaluations use the Claude Code harness and are judged by <code>gpt-5.4-2026-03-05</code>.</li>
<li>SWE-MM: Scores are evaluated on the Claude Code harness using the public dev split of SWE-bench Multimodal, with the modifications described in Appendix 8.3 of the Claude Opus 4.7 system card.</li>
<li>Empty cells (--) indicate that results are not yet available or not applicable.</li>
</ol></div>
</div>
Quickstart
For streamlined integration, we recommend using Qwen3.8 via APIs.
Serving Qwen3.8
[!Important]
Inference efficiency and throughput vary significantly across frameworks.
We recommend using the latest framework versions to ensure optimal performance and compatibility.
For production workloads or high-throughput scenarios, dedicated serving engines such as SGLang, vLLM, or TokenSpeed are recommended.
Qwen3.8 can be deployed with popular inference frameworks, e.g.:
API Usage
[!Important]
Qwen3.8 models operate in thinking mode by default, generating thinking content signified by <think>\n...</think>\n\n before producing the final response.
To disable thinking content and obtain a direct response, refer to the examples here.
[!Tip]
We recommend using the following sets of sampling parameters for generation:
- Thinking Mode:
temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0
- Instruct (or non-thinking) mode:
temperature=0.7, top_p=0.80, top_k=20, min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0
Please note that the support for sampling parameters varies according to inference frameworks.
Qwen3.8 comes with official support for reasoning_effort, which can be used to adjust reasoning depth and control cost:
xhigh (default): for complex tasks demanding thorough analysis
medium: balancing accuracy and speed
low: efficient reasoning optimizing for speed and cost
In addition, preserve_thinking is enabled by default for all workloads for the best out-of-the-box experience. To disable preserved thinking, refer to the examples here.
[!Tip]
In multi-turn agentic tasks, lower reasoning effort does not always reduce overall task completion time. Although it may produce faster per-turn responses, it can also lead to insufficient analysis, more failures, and repeated retries, which may increase total latency and token consumption.
Chat Completions API
The Chat Completions API can be used with most inference frameworks, as well as Qwen Cloud.
Before starting, make sure the OpenAI Python SDK is installed and the API key and the API base URL are configured, e.g.:
pip install -U openai
# Set the following accordingly
export OPENAI_BASE_URL='your-base-url'
export OPENAI_API_KEY='your-api-key'
Text-Only Input
from openai import OpenAI
# Configured by environment variables
client = OpenAI()
messages = [{"role": "user", "content": "Write a Python function to merge two sorted linked lists."}]
completion = client.chat.completions.create(
model="Qwen/Qwen3.8-27B",
messages=messages,
extra_body={
"chat_template_kwargs": {
"enable_thinking": True, # on by default
"preserve_thinking": True, # on by default
},
},
reasoning_effort="xhigh", # xhigh by default; supported levels are xhigh, medium, and low
stream=True,
stream_options={"include_usage": True},
)
reasoning_content = ""
answer_content = ""
is_answering = False
print("\n" + "=" * 20 + "Reasoning" + "=" * 20 + "\n")
for chunk in completion:
if not chunk.choices:
print("\nUsage:")
print(chunk.usage)
continue
delta = chunk.choices[0].delta
if hasattr(delta, "reasoning_content") and delta.reasoning_content is not None:
if not is_answering:
print(delta.reasoning_content, end="", flush=True)
reasoning_content += delta.reasoning_content
elif hasattr(delta, "reasoning") and delta.reasoning is not None:
if not is_answering:
print(delta.reasoning, end="", flush=True)
reasoning_content += delta.reasoning
if hasattr(delta, "content") and delta.content:
if not is_answering:
print("\n" + "=" * 20 + "Answer" + "=" * 20 + "\n")
is_answering = True
print(delta.content, end="", flush=True)
answer_content += delta.content
messages.append({
"role": "assistant",
"content": answer_content,
"reasoning_content": reasoning_content,
"reasoning": reasoning_content,
})
Image Input
from openai import OpenAI
# Configured by environment variables
client = OpenAI()
messages = [
{
"role": "user",
"content": [
{
"type": "image_url",
"image_url": {
"url": "https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.5/demo/CI_Demo/mathv-1327.jpg"
}
},
{
"type": "text",
"text": "The centres of the four illustrated circles are in the corners of the square. The two big circles touch each other and also the two little circles. With which factor do you have to multiply the radii of the little circles to obtain the radius of the big circles?\nChoices:\n(A) $\\frac{2}{9}$\n(B) $\\sqrt{5}$\n(C) $0.8 \\cdot \\pi$\n(D) 2.5\n(E) $1+\\sqrt{2}$"
}
]
}
]
chat_response = client.chat.completions.create(
model="Qwen/Qwen3.8-27B",
messages=messages,
)
print("Chat response:", chat_response)
Video Input
from openai import OpenAI
# Configured by environment variables
client = OpenAI()
messages = [
{
"role": "user",
"content": [
{
"type": "video_url",
"video_url": {
"url": "https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.5/demo/video/N1cdUjctpG8.mp4"
}
},
{
"type": "text",
"text": "How many porcelain jars were discovered in the niches located in the primary chamber of the tomb?"
}
]
}
]
chat_response = client.chat.completions.create(
model="Qwen/Qwen3.8-27B",
messages=messages,
)
# When vLLM is launched with `--media-io-kwargs '{"video": {"num_frames": -1}}'`,
# video frame sampling can be configured via `extra_body` (e.g., by setting `fps`).
# This feature is currently supported only in vLLM.
#
# By default, `fps=2` and `do_sample_frames=True`.
# With `do_sample_frames=True`, you can customize the `fps` value to set your desired video sampling rate.
# chat_response = client.chat.completions.create(
# model="Qwen/Qwen3.8-27B",
# messages=messages,
# extra_body={
# "mm_processor_kwargs": {"fps": 2, "do_sample_frames": True},
# },
# )
print("Chat response:", chat_response)
Instruct (or Non-Thinking) Mode
Qwen3.8-27B will think by default before responding.
You can obtain a direct response from the model without thinking by configuring the API parameters.
For example,
from openai import OpenAI
# Configured by environment variables
client = OpenAI()
messages = [
{
"role": "user",
"content": [
{
"type": "image_url",
"image_url": {
"url": "https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.5/demo/RealWorld/RealWorld-04.png"
}
},
{
"type": "text",
"text": "Where is this?"
}
]
}
]
chat_response = client.chat.completions.create(
model="Qwen/Qwen3.8-27B",
messages=messages,
temperature=0.7,
top_p=0.8,
presence_penalty=1.5,
extra_body={
"top_k": 20,
"chat_template_kwargs": {"enable_thinking": False},
},
)
print("Chat response:", chat_response)
[!Note]
If you are using APIs from Qwen Cloud, in addition to changing model, please use "enable_thinking": False instead of "chat_template_kwargs": {"enable_thinking": False}.
Disable Preserved Thinking
By default, Qwen3.8 retains thinking blocks from all historical messages, maintaining a complete reasoning trace across the conversation. This behavior, known as preserved thinking, ensures full context continuity and is especially beneficial for agent scenarios where decision consistency and reduced redundant reasoning are critical. It also improves KV cache utilization, optimizing inference efficiency in both thinking and non-thinking modes.
If you prefer to retain only the thinking blocks from the latest user message, you can disable this behavior by setting preserve_thinking to False:
from openai import OpenAI
# Configured by environment variables
client = OpenAI()
messages = [...]
chat_response = client.chat.completions.create(
model="Qwen/Qwen3.8-27B",
messages=messages,
extra_body={
"chat_template_kwargs": {"preserve_thinking": False},
},
)
print("Chat response:", chat_response)
[!Note]
If you are using APIs from Qwen Cloud, in addition to changing model, please use "preserve_thinking": False directly instead of wrapping it in chat_template_kwargs.
Best Practices
To achieve optimal performance, we recommend the following settings:
-
Sampling Parameters: We suggest using the following sets of sampling parameters:
- Thinking Mode:
temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0
- Instruct (or non-thinking) mode:
temperature=0.7, top_p=0.80, top_k=20, min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0
For supported frameworks, you can adjust the presence_penalty parameter between 0 and 2 to reduce endless repetition. However, using a higher value may occasionally result in language mixing and a slight decrease in model performance.
-
Adequate Output Length: To optimize performance on agentic tasks, we recommend allocating sufficient output length to allow the model to generate detailed and comprehensive responses. For frameworks that support separate token limits for internal reasoning and final outputs, we suggest the following configuration within the 1M context length:
- Reasoning Content: Set the maximum output length to 262,144 tokens.
- Final Response: Set the maximum output length to 131,072 tokens.
These settings provide the necessary capacity for complex reasoning while ensuring ample space for high-quality final deliverables.
-
Processing Ultra-Long Texts: Qwen3.8-27B natively supports context lengths of up to 262,144 tokens. For long-horizon tasks where the total length (including both input and output) exceeds this limit, we recommend using RoPE scaling techniques to handle long texts effectively, e.g., YaRN.
YaRN is currently supported by several inference frameworks, e.g., vLLM, SGLang, and TokenSpeed.
In general, there are two approaches to enabling YaRN for supported frameworks:
-
Modifying the model configuration file:
In the config.json file, change the rope_parameters fields in text_config to:
{
"mrope_interleaved": true,
"mrope_section": [
11,
11,
10
],
"rope_type": "yarn",
"rope_theta": 10000000,
"partial_rotary_factor": 0.25,
"factor": 4.0,
"original_max_position_embeddings": 262144,
}
-
Passing command line arguments:
For vLLM, you can use
VLLM_ALLOW_LONG_MAX_MODEL_LEN=1 vllm serve ... --hf-overrides '{"text_config": {"rope_parameters": {"mrope_interleaved": true, "mrope_section": [11, 11, 10], "rope_type": "yarn", "rope_theta": 10000000, "partial_rotary_factor": 0.25, "factor": 4.0, "original_max_position_embeddings": 262144}}}' --max-model-len 1000000
For SGLang, you can use
SGLANG_ALLOW_OVERWRITE_LONGER_CONTEXT_LEN=1 python -m sglang.launch_server ... --json-model-override-args '{"text_config": {"rope_parameters": {"mrope_interleaved": true, "mrope_section": [11, 11, 10], "rope_type": "yarn", "rope_theta": 10000000, "partial_rotary_factor": 0.25, "factor": 4.0, "original_max_position_embeddings": 262144}}}' --context-length 1000000
For TokenSpeed, you can use
TOKENSPEED_ALLOW_OVERWRITE_LONGER_CONTEXT_LEN=1 tokenspeed serve ... --hf-overrides '{"text_config": {"rope_parameters": {"mrope_interleaved": true, "mrope_section": [11, 11, 10], "rope_type": "yarn", "rope_theta": 10000000, "partial_rotary_factor": 0.25, "factor": 4.0, "original_max_position_embeddings": 262144}}}' --max-model-len 1000000
[!NOTE]
All the notable open-source frameworks implement static YaRN, which means the scaling factor remains constant regardless of input length, potentially impacting performance on shorter texts.
We advise modifying the rope_parameters configuration only when processing long contexts is required.
It is also recommended to modify the factor as needed. For example, if the typical context length for your application is 524,288 tokens, it would be better to set factor as 2.0.
-
Long Video Understanding: To optimize inference efficiency for plain text and images, the size parameter in the released video_preprocessor_config.json is conservatively configured. It is recommended to set the longest_edge parameter in the video_preprocessor_config file to 469,762,048 (corresponding to 224k video tokens) to enable higher frame-rate sampling for hour-scale videos and thereby achieve superior performance. For example,
{"longest_edge": 469762048, "shortest_edge": 4096}
Alternatively, override the default values via engine startup parameters. For implementation details, refer to: vLLM / SGLang.
Citation
If you find our work helpful, feel free to give us a cite.
@misc{qwen38,
title = {{Qwen3.8-Max}: A New Bar for Coding and Cowork},
url = {https://qwen.ai/blog?id=qwen3.8},
author = {{Qwen Team}},
month = {August},
year = {2026}
}
<h2>EXAMPLE GENERATION(S):</h2>
Q4KS, non imatrix, standard Qwen settings, NO cache compression of any kind.
NOTE: Some formatting may be lost on copy/paste/export.
<style type="text/css">
@page { size: 21cm 29.7cm; margin: 2cm }
p { line-height: 115%; margin-bottom: 0.25cm; background: transparent }
h1 { margin-bottom: 0.21cm; background: transparent; page-break-after: avoid }
h1.western { font-family: "Liberation Serif", serif; font-weight: bold; font-size: 24pt }
h1.cjk { font-size: 24pt; font-family: "NSimSun"; font-weight: bold }
h1.ctl { font-family: "Arial"; font-size: 24pt; font-weight: bold }
h2 { margin-top: 0.35cm; margin-bottom: 0.21cm; background: transparent; page-break-after: avoid }
h2.western { font-family: "Liberation Serif", serif; font-weight: bold; font-size: 18pt }
h2.cjk { font-size: 18pt; font-family: "NSimSun"; font-weight: bold }
h2.ctl { font-family: "Arial"; font-size: 18pt; font-weight: bold }
td p { orphans: 0; widows: 0; background: transparent }
th p { font-weight: bold; text-align: center; orphans: 0; widows: 0; background: transparent }
strong { font-weight: bold }
em { font-style: italic }
a:link { color: #000080; text-decoration: underline }
a:visited { color: #800000; text-decoration: underline }
</style>
<br/>
Example #1 of 4, using “xhigh” (default reasoning)</p>
<p style="line-height: 100%; margin-bottom: 0cm"><br/>
</p>
<p style="line-height: 100%; margin-bottom: 0cm">This is about 1/3 to
½ the size of normal output.<br/>
<br/>
Example #2, same prompt
[below], but at “medium”.</p>
<p style="line-height: 100%; margin-bottom: 0cm"><br/>
</p>
<p style="line-height: 100%; margin-bottom: 0cm">---</p>
<p style="line-height: 100%; margin-bottom: 0cm"><br/>
</p>
<p style="line-height: 100%; margin-bottom: 0cm"><b>PROMPT:<br/>
<br/>
Explain
10 ways (in extensive detail) to use the "night" time
cooling of radiant energy into space to reduce global temperatures.</b></p>
<p style="line-height: 100%; margin-bottom: 0cm"><br/>
</p>
<p><span style="display: inline-block; border: none; padding: 0cm"><font color="#c9211e"><b>Thought
for 1 minutes 28 seconds</b></font></span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">We
need answer user: "Explain 10 ways (in extensive detail) to use
the 'night' time cooling of radiant energy into space to reduce
global temperatures." Need interpret: nighttime cooling via
thermal radiation to space, passive radiative cooling, perhaps
enhancement. Need give 10 ways in extensive detail. We need ensure
scientifically accurate, maybe distinguish passive radiative cooling,
selective emitters, sky windows, materials, geometry, location,
systems, urban/industrial/agricultural, refrigeration, power, carbon
capture, etc. Need discuss feasibility, mechanisms, caveats: Earth
emits in mid-IR; atmosphere has window 8-13 microns; radiative
cooling to sky possible when surface/emitter cooler than effective
radiating temp, requires selective spectral emissivity high in sky
window and low/reflective elsewhere to minimize solar/IR absorption;
works best clear nights, low humidity, low wind, high altitude. We
should not overclaim. Need 10 ways maybe: (1) passive radiative
cooling materials for buildings; (2) architectural coatings and
membranes; (3) nighttime cooling towers/heat sinks for data
centers/industrial; (4) radiative chillers/refrigeration; (5)
district cooling networks with radiative coolers; (6) agricultural
cold storage and evaporative/condensing? Maybe use to cool water? (7)
enhanced space radiative cooling with selective emitters at large
scale (engineered surfaces or films) to increase planetary
albedo/emissivity? (8) nighttime thermal storage: cool materials by
night, store thermal energy/cold for daytime; (9) integration with
solar-thermal or solar-electric plants: nighttime radiative cooling
of heat exchangers to increase efficiency and reduce thermal
pollution; (10) urban heat island mitigation and microclimate (cool
pavements, roofs, facades); maybe "cooling the land via
engineered materials in arid regions". Also (maybe) "radiative
cooling of atmospheric water vapor / condensation"? Use
condensation? Or "cooling the ocean?" Might be too
speculative.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Need
explain "night time cooling of radiant energy into space"
perhaps as a method: increase emission of infrared radiation to outer
space during nights, lowering surface temperature. Need 10 ways in
detail. We can organize.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Potential
ways:</span></p>
<ol>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Passive
radiative cooling coatings for roofs/walls: use selective emitters
(e.g., BaSO4, polymer films with micro/nanostructures, multilayer
dielectric films) high emissivity in 8-13 μm sky window and high
solar reflectance; reduces roof temp, indoor cooling load; deployed
globally in cities.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Large-area
radiative cooling membranes over buildings/infrastructure: tents,
awnings, cool roofs with air gaps; airflow, ventilation; cooling
effect at night and day; can reduce urban heat island.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Night-time
radiative chillers for refrigeration/air conditioning: emitter
panels with heat exchangers and fans; cool water or air below
ambient using sky as heat sink; integrated with evaporative or
desiccant dehumidification; data centers, hospitals, cold storage.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative
cooling for district cooling and industrial process cooling:
centralized rooftop/field radiative coolers that transfer process
waste heat to sky; paired with heat pumps or absorption chillers;
reduce fossil-fuel cooling.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Thermal
energy storage: store "cold" or pre-cooled materials/phase
change materials at night via radiative cooling; use for daytime
cooling; combined with PCM, latent heat; reduce peak load and AC
energy.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Enhance
planetary radiative cooling with engineered surfaces/land cover: in
high-altitude arid regions, deploy large fields of selective
emitters or reflective white surfaces that emit in atmospheric
window to increase outgoing longwave radiation; evaluate albedo,
ecosystem, hydrology impacts.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Cool
pavements and transportation infrastructure: radiatively cooling
asphalt/concrete substitutes, road surfaces, parking lots; reduces
heat stored in pavement and lowers vehicle fuel/AC consumption;
especially in sunny, clear-sky areas.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative
cooling for agriculture: cooling seed stores, livestock housing,
greenhouses, vineyards/orchards; night cooling to prevent heat
stress, reduce spoilage; using selective emitters on roofs or sheds;
can also manage condensation.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Solar
power and solar thermal efficiency: cool PV panels by night/day
using radiative surfaces (backsheet, coatings) and heat pipes;
reduces module temperature and increases efficiency; cool heat
exchangers in solar thermal; maybe reduce thermal stratification.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Data
centers and electronics cooling: dedicated radiative cooling panels
as low-energy heat rejection for servers; nocturnal operation with
fans and heat exchangers; reduces PUE and water use. Alternative:
"Cooling seawater for carbon capture/thermal storage"?
Need 10. Could include "Cooling oceans via selective emitters?"
not realistic. Could include "use to assist
condensation/humidification?" We can mention radiative cooling
can condense moisture; but user asks reduce global temperatures.
Need focus.</span></p></li>
</ol>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Need
maybe mention fundamental physics:</span></p>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Earth
radiates ~390 W/m²; sky emits ~200 W/m²; net radiative cooling
depends on temperature difference; cooling power = εσ(T_s^4 -
T_sky^4) minus absorbed IR and convective; sky window 8-13 μm.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">To
maximize: high emissivity in 8-13 μm, high solar reflectance
(10-2500 nm), minimize convective/absorbed heat; clear skies, high
altitude, low humidity; night operation best.</span></p></li>
</ul>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Need
"extensive detail" maybe each way with mechanisms, design,
implementation, quantification examples, challenges.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Need
ensure not misleading: "reduce global temperatures" not
instant; local and system-level; global effect if large-scale
adoption; could be a small fraction. We can say can contribute to
demand-side mitigation, reducing fossil fuel emissions from cooling,
but direct global cooling depends on large area.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Need
maybe include "night-time" specific: all methods can be
operated/optimized at night. Some are passive (continuous). Could
mention:</span></p>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Passive
roofs: cool at night via radiative cooling, lower indoor temp.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative
cooling towers: at night, radiative cooling of water below ambient;
day still useful if selective and reflective, but night better.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Solar-thermal:
store heat in high temp? Wait use radiative cooling to reject waste
heat and increase cycle efficiency, but not directly lower ambient.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Use
"night-time radiative cooling as an auxiliary refrigerant"
by cooling heat exchanger fluid, driving absorption chiller.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Use
"radiative cooling to produce cold for ice-making" in
developing regions.</span></p></li>
</ul>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Potential
10 in more nuanced way:</span></p>
<ol>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Passive
radiative cool roofs with selective emitters.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Active
radiative cooling systems (fan-assisted) for building HVAC.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative
chillers and absorption/heat-pump assisted cooling for
process/refrigeration.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">District
cooling networks using centralized radiative heat rejection.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Night-time
cold storage and phase-change thermal storage.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Large-scale
landscape-level radiative cooling deployments (engineered emitters)
in suitable regions.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative-cooled
pavements/roads/infrastructure for urban heat island mitigation.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Agricultural
cooling (greenhouses, livestock, cold chain) using radiative
cooling.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative
cooling of solar PV/thermal collectors to improve efficiency and
reduce thermal load.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Data
center/electronics radiative cooling and night-time heat rejection.</span></p></li>
</ol>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Need
perhaps separate "radiative cooling membranes/films" and
"coatings". Could combine. Need ensure 10 distinct.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Maybe
include "cooling water bodies/evaporation" or "radiative
cooling to enhance ocean-atmosphere exchange"? Hmm.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Let's
design final answer: intro with physics. Then 10 sections. For each:
concept, how it works, detailed implementation, quantitative
potential, key challenges.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Need
answer not too short. Oververbosity 9: thorough. Need be careful with
scientific terms.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Let's
draft in mind.</span></p>
<p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Intro:</span></p>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative
cooling: any body with T > environment emits thermal radiation;
if emissivity high in atmospheric transparency window (8-13 μm), it
can radiate to cold outer space; at night, effective sky temp ~-40
to -70 C or radiative heat flux can be 50-150 W/m² under clear
skies; selective surfaces can be 10-30 C below ambient.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Requirements:
high IR emissivity in 8-13 μm; high reflectance in solar 300-2500
nm; low thermal conductance or airflow; clear, dry, high-altitude
climates best.</span></p></li>
</ul>
<ol>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Passive
radiative cool roofs/facades.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Materials:
multilayer dielectric films, BaSO4/polymer, micro-nanostructured
photonic crystals, ceramic composites.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Mechanism:
emits in sky window, reflects sunlight; reduces roof temp by 10-30
C; reduces HVAC.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Deployment:
roofing membranes, tiles, paints, cool glass; cost-effective; reduce
building energy demand and urban heat.</span></p></li>
</ul>
<ol start="2">
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative-cooled
building envelopes and membranes with airflow.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Use
panels/tarps with air gaps, louvers, night vents; passive stack
effect; night purging.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Mechanism:
cool shell at night; daytime insulation; lower indoor temp;
especially hot-arid.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Detailed:
double-skin roofs, ventilated facades, cool awnings; sensors
controlling vents; reduce peak heat gains.</span></p></li>
</ul>
<ol start="3">
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Active
radiative chillers for air conditioning/refrigeration.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Components:
selective emitter, heat exchanger, water/air circuit, fans,
sometimes heat pump/absorption.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Mechanism:
pump warm fluid from indoor coils to emitter; emitter radiates heat
to sky; can cool fluid below ambient at night; drive AC or store
cold.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Example:
roof-top units; 20-50 W/m² cooling at night; can offset 50-100% AC
in mild climates.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Integration:
desiccant dehumidification to manage humidity; hybrid with
evaporative cooling.</span></p></li>
</ul>
<ol start="4">
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">District
cooling and industrial waste-heat rejection via radiative coolers.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Central
plant radiative cooling fields (emitter arrays, cooling towers) at
night; connect to district cooling loops; industrial process cooling
(plastics, food, chemical, data centers).</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Use
low-grade heat; reduce chiller electricity; night operation shifts
load.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Large
area: thousands of m²; use in sunny arid cities; potential kW-MW.</span></p></li>
</ul>
<ol start="5">
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Night-time
cold storage and phase-change thermal storage.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Cool
PCM or water at night via radiative emitters; store thermal energy
as cold; release next day.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Example:
paraffin/eutectic salts at 5-25 C; insulated tanks with radiative
coolers; provide pre-cooling for buildings, cold rooms.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Use
for peak shaving and resilience; reduces AC runtime and grid
emissions.</span></p></li>
</ul>
<ol start="6">
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Large-scale
land-based radiative cooling deployments.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Install
selective emitter arrays on rooftops, deserts, high plateaus;
increase outgoing longwave radiation through atmospheric window.</span></p></li>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Mechanism:
more of surface heat emitted to space; lower surface/albedo? Need
reflect solar too.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Challenges:
land use, ecosystem, water, wind; requires governance; can reduce
urban heat island and local warming; global effect small but could
matter with millions km².</span></p></li>
</ul>
<ol start="7">
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative-cooled
pavements/roads and transport infrastructure.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Use
cool pavement coatings, aggregates, membranes with high 8-13 μm
emissivity and high solar reflectance; cool roads/parking lots;
reduce heat stored and reradiated; lower vehicle fuel/AC; reduce
UHI.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Deployment:
parking lots, highways, airport aprons; use in hot regions;
integrate with drainage.</span></p></li>
</ul>
<ol start="8">
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Agricultural
cooling and cold chain.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Use
radiative cooling shelters for livestock, poultry, greenhouses,
produce storage; night cooling of barns; reduce heat stress,
mortality, spoilage; lower refrigeration energy.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Details:
insulated barn roofs with emitter panels; cold storage warehouses
with radiative chiller; dew management; integration with
dehumidification.</span></p></li>
</ul>
<ol start="9">
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Radiative
cooling of solar PV/thermal systems.</span></p></li>
</ol>
<ul>
<li><p style="margin-bottom: 0cm; border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Use
cool backsheets, selective emitter panels behind PV, heat pipes;
cool modules by day/night; reduce PV cell temperature; improve
efficiency; for solar-thermal, use radiative coolers as night-time
heat sinks to recover energy and reduce thermal losses.</span></p></li>
<li><p style="border: none; padding: 0cm"><span style="display: inline-block; border: none; padding: 0cm">Could
include concentrating solar: radiatively cool mirrors/absorbers at
The model card is truncated. Read the rest on Hugging Face.