Downloads · 30 days
256
6% of all-time downloads
jashepp/Agents-A1-35B-A3B-MXFP4_MOE_Hybrid-Imatrix-GGUF
Agents-A1-35B-A3B-MXFP4_MOE_Hybrid-Imatrix-GGUF is a text generation model from jashepp. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
Agents‑A1 is a 35B Mixture-of-Experts agentic model built to scale heterogeneous agent abilities across long-horizon search, engineering, scientific research, instruction following, and tool calling.
Downloads · 30 days
256
6% of all-time downloads
All-time downloads
4K
Public
Repo size
59.2 GB
Likes
2
Public
Click a slice to open those files.
.gguf59.2 GB · 100%
From the Hugging Face model README
Agents‑A1 is a 35B Mixture-of-Experts agentic model built to scale heterogeneous agent abilities across long-horizon search, engineering, scientific research, instruction following, and tool calling.
This repository contains custom, highly optimized, multi-tier mixed precision GGUF weights for InternScience/Agents-A1.
<img style="width:100%;max-width:900px;" src="https://huggingface.co/InternScience/Agents-A1/resolve/main/figures/a1_benchmarks_altair_grid.svg" alt="Agents-A1 Benchmark Overview">[!TIP] ℹ️ Agents‑A1 is a reasoning/thinking model. It works well for advanced agentic and programming tasks.
However I still recommend using Ornith-1.0-35B for faster performance & better quality results.
To learn more about Agents‑A1, read the homepage.
To learn more about how to use Agents‑A1, view the base model.
These quants were generated using manual layer targeting to maximize quality while shrinking the massive VRAM footprint of the Mixture of Experts layers.
In order of quality:
| Filename | Size | Quants |
|---|---|---|
| Agents-A1-35B-A3B-MXFP4_MOE_Q8_0_F16-Imatrix.gguf | 20.7 GB | MXFP4_MOE + Q8_0 + F16 |
| Agents-A1-35B-A3B-MXFP4_MOE_Q8_0-Imatrix.gguf | 19.8 GB | MXFP4_MOE + Q8_0 |
| Agents-A1-35B-A3B-MXFP4_MOE-Only-Imatrix.gguf | 18.5 GB | MXFP4_MOE Only |
The following datasets were used for the imatrix:
I am currently remaking this imatrix with 2048 ctx, 3000 chunks, with smaller sets of text
Old/Current: Custom Target Matrix (2048 ctx with 600 chunks), up to a max of 100MB of each:
Standard global quantization presets (like stock MXFP4_MOE) compress the backbone layers uniformly, which degrades the delicate reasoning capabilities of advanced agent models.
This repository provides multiple distinct manual configuration layouts to balance precision and memory constraints:
MXFP4 + Q8_0 + F16)Agents-A1-35B-A3B-MXFP4_MOE_Q8_0_F16-Imatrix.gguf - Designed for maximum quality preservation, this layout implements a strict 3-Tier Precision Matrix:
token_embd.weight, output.weight - Protects the critical input/output vocabulary mappings. Dramatically prevents text degradation.ssm_alpha, ssm_beta - Protects the integrity of the Mamba state-space calculations across long-range context tokens.ssm_out, *._shexp - Keeps the attention mechanics, and all trailing shared experts at high quality, to protect the logical research loops.ffn_down_exps, ffn_gate_exps, ffn_up_exps - Shrink the massive background expert parameters directly to MXFP4.MXFP4 + Q8_0)Agents-A1-35B-A3B-MXFP4_MOE_Q8_0-Imatrix.gguf - Designed for a slightly leaner memory profile, this layout utilizes 2-Tier Precision:
Q8_0 format.MXFP4.Agents-A1-35B-A3B-MXFP4_MOE-Only-Imatrix.gguf - Using only MXFP4, this shrinks the model down to 18.5 GB. The quality is not the best, but it can still do decent work.
MXFP4, for speed and a low VRAM footprint.These files were converted via llama-quantize utilizing the following manual recipe parameters:
# Requires python3.12, with `pip install --upgrade transformers`
python convert_hf_to_gguf.py "Ornith-1.0-35B/" --outtype f16 --outfile "Ornith-1.0-35B_F16.gguf"
llama-quantize \
--tensor-type ".*_shexp\.weight=Q8_0" \
--tensor-type "token_embd\.weight=F16" \
--tensor-type "^output\.weight=F16" \
--tensor-type "blk\..*\.(ssm_alpha|ssm_beta)\.weight=F16" \
--tensor-type "blk\..*\.(ffn_down_exps|ffn_gate_exps|ffn_up_exps)\.weight=MXFP4" \
--imatrix "imatrix.gguf" \
"Agents-A1-35B-A3B_F16.gguf" \
"Agents-A1-35B-A3B-MXFP4_MOE_Q8_0_F16-Imatrix.gguf" \
Q8_0
llama-quantize \
--tensor-type ".*_shexp\.weight=Q8_0" \
--tensor-type "blk\..*\.(ffn_down_exps|ffn_gate_exps|ffn_up_exps)\.weight=MXFP4" \
--imatrix "imatrix.gguf" \
"Agents-A1-35B-A3B_F16.gguf" \
"Agents-A1-35B-A3B-MXFP4_MOE_Q8_0-Imatrix.gguf" \
Q8_0
llama-quantize \
--tensor-type ".*_shexp\.weight=MXFP4" \
--tensor-type "token_embd\.weight=MXFP4" \
--tensor-type "^output\.weight=MXFP4" \
--tensor-type "blk\..*\.(ssm_alpha|ssm_beta|ssm_out|attn_gate|attn_qkv|ffn_down|ffn_gate|ffn_up|attn_k|attn_q|attn_v|attn_output)\.weight=MXFP4" \
--tensor-type "blk\..*\.(ffn_down_exps|ffn_gate_exps|ffn_up_exps)\.weight=MXFP4" \
--imatrix "imatrix.gguf" \
"Agents-A1-35B-A3B_F16.gguf" \
"Agents-A1-35B-A3B-MXFP4_MOE-Only-Imatrix.gguf" \
MXFP4_MOE
llama-server Configuration (config.ini)To maintain the rock-solid reasoning loop and prevent agents from falling into repetitive tool-calling deadlocks, use the following server parameter recommendations (Similar to other Qwen 3.5+ configurations).
# --- Samplers (Dynamic & Expressive) ---
temperature = 0.55
top-k = 15
top-p = 0.90
min-p = 0.15
# --- Penalties (Prevent Syntax & Reasoner Corruption) ---
repeat-penalty = 1.08
presence-penalty = 0.00
# --- DRY Sampler (Protects Indentation & Structural Boilerplate) ---
dry-multiplier = 0.8
dry-base = 1.75
dry-allowed-length = 5
dry-penalty-last-n = 1024
dry-sequence-breaker = ["\n", ":", " ", "\t", "\"", ","]
# --- Enforced Execution Graph ---
samplers = min_p;top_k;top_p;temp;dry
Highly Recommended: Always keep reasoning/thinking enabled, for better quality results.
# --- Reasoning ---
chat-template-kwargs = { "enable_thinking":true }
reasoning = on
reasoning-budget = 4096
reasoning-format = auto
This works well with 256k context window.
[!TIP] For Maximum Quality at 100k+ Context:
Use the MXFP4_MOE + Q8_0 + F16 split-quantized version.
- Preserved at F16:
token_embd.weight,output.weight,*.ssm_alpha.weight, and*.ssm_beta.weight.- Why this matters: Keeping these critical layers at full precision prevents the model from dropping fine details during extreme "needle-in-a-haystack" retrieval tasks (large context windows).
- What to avoid: If
output.weightor the embedding layers are quantized toQ8_0or lower, logit precision rounds off, causing the model to lose accuracy and forget specific details in long-context scenarios.
Updated 2026-07-08:
I'm doing this as a side hobby, with my AMD 5900X, 64GB DDR4, RTX 3060 12GB & RTX 5060 Ti 16GB.
In addition to the above configuration, I also use:
slots = 1
parallel = 1
no-warmup = true
flash-attn = on
mlock = false
no-mmap = false
context-shift = false
batch-size = 2048
ubatch-size = 256
fit = on
fit-target = 768
main-gpu = 0
split-mode = layer
n-gpu-layers = 999
n-cpu-moe = 0
tensor-split = 16,12
override-tensor = (token_embd)=CUDA0,(vision|vpm|nextn)=CPU
cache-type-k = q8_0
cache-type-v = q8_0
jinja = true
chat-template = jinja
chat-template-file = chat_template.jinja
For further quality and better ssm behaviour, this configuration can help:
context-shift = false
cache-type-k = f16
cache-type-v = f16
As a passionate developer, I'm always programming, automating, or experimenting with new ideas.
I love building open-source tools, trying out new web tech, and creating things that don't yet exist, including local AI & quantizing models.
I love sharing these creations to give back to the community.
If my projects have saved you time or helped you out, consider supporting my work below!
Agents-A1 base model.Released under Apache-2.0.
@misc{bai2026scalinghorizonparametersreaching,
title={Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent},
author={Lei Bai and Zongsheng Cao and Yang Chen and Zhiyao Cui and Shangheng Du and Yue Fan and Shiyang Feng and Zijie Guo and Haonan He and Liang He and Xiaohan He and Shuyue Hu and Yusong Hu and Songtao Huang and Yichen Jiang and Hao Li and Xin Li and Dahua Lin and Weihao Lin and Fenghua Ling and Dongrui Liu and Zhuo Liu and Runmin Ma and Chunjiang Mu and Haoyang Peng and Tianshuo Peng and Jinxin Shi and Luohe Shi and Boyuan Sun and Zelin Tan and Shengji Tang and Qianyi Wang and Yiming Wu and Yi Xie and Xiangchao Yan and Jingqi Ye and Peng Ye and Fangchen Yu and Jiakang Yuan and Bihao Zhan and Bo Zhang and Chen Zhang and Shufei Zhang and Shuaiyu Zhang and Wenlong Zhang and Yiqun Zhang and Junpeng Zhao and Zhijie Zhong and Bowen Zhou and Yuhao Zhou},
year={2026},
eprint={2606.30616},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2606.30616},
}