BigBang1.0-35B-A3B-NPU2

BigBang1.0-35B-A3B-NPU2 (FastFlowLM / Lemonade NPU2 Quantization)

[!IMPORTANT] Quantization & NPU Compatibility Note: This repository contains Q4NX quantized weights converted from endless-frontier/BigBang-v1 to run natively on FastFlowLM (flm) v1.0.3+ and Lemonade on AMD XDNA NPU hardware.

• Model Type: Quantized model conversion (NPU Q4NX format)
• Parent / Base Model: endless-frontier/BigBang-v1
• Details: Re-quantized and packaged for FastFlowLM v1.0.3+ and Lemonade with NPU2 execution plan, pinned 32k context, and multi-turn EOS stop token IDs ([248044, 248046]).
• Architecture: BigBang-v1 (Qwen3.6 MoE 35B-A3B base)
• Quantization Format: Q4_K_M / Q4NX
• Format: Q4NX (safetensors format with AMD NPU block packing). Note that this is not a standard GGUF file; it is executed natively via flm / Lemonade on AMD Ryzen AI NPUs.

---

Serving with Lemonade & FastFlowLM

To serve this model via Lemonade or FastFlowLM:

# Pull and run with FLM:
flm pull BigBang1.0-35B-A3B-NPU2
flm serve BigBang1.0-35B-A3B-NPU2 --ctx-len 32768 --port 8001

Or configure via Lemonade:

lemonade run BigBang1.0-35B-A3B-NPU2

---

Original Model Information (endless-frontier/BigBang-v1)

Below is the model card from the upstream repository endless-frontier/BigBang-v1:

---

BigBang-V1

Introduction

As Large Language Models (LLMs) approach human expert performance, their continued development is increasingly constrained by training tasks conceived within the limits of human knowledge. We argue that open-ended capability growth requires verifiable frontier tasks: problems at the boundary of current knowledge whose solutions can be objectively evaluated through formal methods, computation, simulation, or domain-specific tools. To this end, we introduce BigBang, a general-purpose LLM evolved from Qwen 3.6 35B-A3B through efficient post-training with an adversarial, self-evolving synthetic data framework. The framework contains two core components: (1) generator agents that continually propose and solve increasingly challenging scientific and technical problems, and (2) critic agents that evaluate correctness, difficulty, scalability, and diversity, while using held-out real research tasks to calibrate the evolving synthetic-data distribution. Through iterative generator–critic interaction, the framework constructs approximately 10,000 high-difficulty post-training examples across multiple domains. Despite the modest data scale, BigBang substantially outperforms its base model across scientific research, reasoning, coding, and tool-use benchmarks, achieving aggregate performance between DeepSeek V4 Flash (284B) and DeepSeek V4 Pro (1.6T). These results demonstrate that self-evolving synthesis of verifiable frontier tasks provides a promising path toward scalable and open-ended intelligence.

Main Results

BigBang-V1 on eight representative benchmarks spanning long-horizon search, software engineering, scientific research, and AI research. BigBang-V1 obtains the highest reported score among the selected 35B models on all eight benchmarks. It even exceeds DeepSeek V4 Pro Preview (1.6T) on FrontierScience Research, Humanity's Last Exam, PaperBench(Code-Dev) and BioMysteryBench-HD.

<p align="center"> <a href="./bigbang_main_figure.pdf"> <img src="./bigbang_main_figure.png" alt="BigBang-V1 on eight representative benchmarks" width="100%"> </a> </p> <p align="center"><em>BigBang-V1 on eight representative benchmarks. Click the figure to view the PDF.</em></p>

Benchmark Results

Comparison of BigBang-V1 with representative closed- and open-source frontier models, together with models at the 35B scale, across benchmarks for long-horizon search, coding, scientific research, and AI research. The "-" indicates the score is not publicly available or not tested.

Quickstart

For streamlined integration, we recommend using BigBang-V1 via APIs. Below is a guide to use BigBang-V1 via OpenAI-compatible API.

Serving BigBang-V1

BigBang-V1 can be served via APIs with popular inference frameworks. In the following, we show example commands to launch OpenAI-Compatible API servers for BigBang-V1 models.

[!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, KTransformers or vLLM are strongly recommended.

[!Important] The model has a default context length of 262,144 tokens. If you encounter out-of-memory (OOM) errors, consider reducing the context window. However, because BigBang-V1 leverages extended context for complex tasks, we advise maintaining a context length of at least 128K tokens to preserve thinking capabilities.

SGLang

SGLang is a fast serving framework for large language models and vision language models. sglang>=0.5.10 is recommended for BigBang-V1, which can be installed using the following command in a fresh environment:

uv pip install sglang[all]

See its documentation for more details.

The following will create API endpoints at http://localhost:8000/v1:

• Standard Version: The following command can be used to create an API endpoint with maximum context length 262,144 tokens using tensor parallel on 8 GPUs.

  python -m sglang.launch_server --model-path endless-frontier/BigBang-v1 --port 8000 --tp-size 8 --mem-fraction-static 0.8 --context-length 262144 --reasoning-parser qwen3
  

• Tool Use: To support tool use, you can use the following command.

  python -m sglang.launch_server --model-path endless-frontier/BigBang-v1 --port 8000 --tp-size 8 --mem-fraction-static 0.8 --context-length 262144 --reasoning-parser qwen3 --tool-call-parser qwen3_coder
  

• Multi-Token Prediction (MTP): The following command is recommended for MTP:

  python -m sglang.launch_server --model-path endless-frontier/BigBang-v1 --port 8000 --tp-size 8 --mem-fraction-static 0.8 --context-length 262144 --reasoning-parser qwen3 --speculative-algo NEXTN --speculative-num-steps 3 --speculative-eagle-topk 1 --speculative-num-draft-tokens 4
  

For detailed deployment guide, see the SGLang Qwen3.5 Cookbook.

vLLM

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. vllm>=0.19.0 is recommended for BigBang-V1, which can be installed using the following command in a fresh environment:

uv pip install vllm --torch-backend=auto

See its documentation for more details.

The following will create API endpoints at http://localhost:8000/v1:

• Standard Version: The following command can be used to create an API endpoint with maximum context length 262,144 tokens using tensor parallel on 8 GPUs.

  vllm serve endless-frontier/BigBang-v1 --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3 
  

• Tool Call: To support tool use, you can use the following command.

  vllm serve endless-frontier/BigBang-v1 --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder 
  

• Multi-Token Prediction (MTP): The following command is recommended for MTP:

  vllm serve endless-frontier/BigBang-v1 --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3 --speculative-config '{"method":"qwen3_next_mtp","num_speculative_tokens":2}'
  

• Text-Only: The following command skips the vision encoder and multimodal profiling to free up memory for additional KV cache:

  vllm serve endless-frontier/BigBang-v1 --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3 --language-model-only
  

For detailed deployment guide, see the vLLM Qwen3.5 Recipe.

KTransformers

KTransformers is a flexible framework for experiencing cutting-edge LLM inference optimizations with CPU-GPU heterogeneous computing. For running BigBang-V1 with KTransformers, see the KTransformers Deployment Guide.

Hugging Face Transformers

Hugging Face Transformers contains a lightweight server which can be used for quick testing and moderate load deployment. The latest transformers is required for BigBang-V1:

pip install "transformers[serving]"

See its documentation for more details. Please also make sure torchvision and pillow are installed.

Then, run transformers serve to launch a server with API endpoints at http://localhost:8000/v1; it will place the model on accelerators if available:

transformers serve endless-frontier/BigBang-v1 --port 8000 --continuous-batching