nota-ai/Solar-Open2-250B-Nota-NVFP4 · Hugging Face

Source: https://huggingface.co/nota-ai/Solar-Open2-250B-Nota-NVFP4 Nota AIpresents a 4-bit quantized release ofUpstage’sSolar Open2 250B, produced with Nota AI’s proprietary quantization technology specialized for Mixture-of-Experts (MoE) large language models.

https://huggingface.co/nota-ai/Solar-Open2-250B-Nota-NVFP4#highlightsHighlights

• NVFP4 (4-bit float, W4A4)—group\_size=16, packed in thellm\-compressor(compressed-tensors) format for direct serving invLLM. Both weights and activations are quantized to 4-bit floating point.- > **Requires NVIDIA Blackwell.**NVFP4 relies on the FP4 tensor cores introduced in the Blackwell architecture (e.g. B200 / GB200), so inference must run on a Blackwell-class GPU. Earlier architectures (Hopper, Ada, Ampere) do not support NVFP4 execution.
• Nota AI’s proprietary MoE quantization framework.This release is built upon a suite of techniques developed by Nota AI to preserve model quality under aggressive low-bit quantization of MoE architectures:- A MoE-specialized calibration-dataset construction method, which achieved1st place across all tracksat the NVIDIA Nemotron Hackathon. - DREAM-MoEandSRA-MoE, two quantization algorithms proposed by Nota AI (published at theICML 2026 Workshop on AdaptFM), which preserve MoE routing decisions and align expert-routing behavior throughout the quantization process.

https://huggingface.co/nota-ai/Solar-Open2-250B-Nota-NVFP4#licenseLicense

Solar Open 2 is distributed under theUpstage Solar License.

Key requirements for Derivative AI Models(create / train / fine-tune / distill / improve using Solar Open 2):

• **Naming:**prefix your model name with “Solar” (e.g.,Solar\-MyModel\-v1).
• **Attribution:**prominently display “Built with Solar” in related public-facing materials.
• **Notice:**include a copy of the Upstage Solar License with your derivative model.

https://huggingface.co/nota-ai/Solar-Open2-250B-Nota-NVFP4#performancePerformance

https://huggingface.co/nota-ai/Solar-Open2-250B-Nota-NVFP4#weight-footprintWeight footprint

PrecisionWeight footprintBF16500.6 GBNota NVFP4153.3 GB

https://huggingface.co/nota-ai/Solar-Open2-250B-Nota-NVFP4#benchmarksBenchmarks

BenchmarkBF16Nota NVFP4Tau2-Bench75.2075.08HLE27.8827.66GPQA Diamond86.2685.45IFBench80.0081.02LiveCodeBench (v5–v6)87.0388.55MMLU-Pro86.1986.15AIME 2026 (EN)95.6796.67IFEval (EN)94.0992.61HMMT92.0590.15KMMLU-Pro78.3877.93HAE-RAE Bench v1.173.8472.84AIME (KO)97.6797.00KBL75.5175.40KBank-MMLU80.8080.68KorMedMCQA92.9993.05Avg.81.5781.35

https://huggingface.co/nota-ai/Solar-Open2-250B-Nota-NVFP4#quick-startQuick Start

This model is packed in the NVFP4 (compressed-tensors) format and can be served directly with vLLM on aBlackwell-class GPU:

uv venv --python 3.12 --seed solar_open2_venv
source .venv/bin/activate

VLLM_PRECOMPILED_WHEEL_LOCATION="https://github.com/vllm-project/vllm/releases/download/v0.22.0/vllm-0.22.0%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl" \
VLLM_USE_PRECOMPILED=1 \
uv pip install --reinstall-package vllm --torch-backend=cu129 \
  "git+https://github.com/UpstageAI/[email protected]"

vllm serve nota-ai/Solar-Open2-250B-Nota-NVFP4 \
  --served-model-name solar-open2-250b \
  --tensor-parallel-size 4 \
  --default-chat-template-kwargs '{"think_render_option":"preserved"}' \
  --reasoning-parser solar_open2 \
  --tool-call-parser solar_open2 \
  --enable-auto-tool-choice \
  --logits-processors vllm.v1.sample.logits_processor.solar_open2:SolarOpen2TemplateLogitsProcessor

• Set\-\-tensor\-parallel\-sizeaccording to the number of GPUs available in your serving environment.
• See the original model card for the prompt format, parser configuration, and further details.

Send a chat completion request:

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "solar-open2-250b",
    "messages": [
      {"role": "user", "content": "What is Upstage?"}
    ],
    "max_tokens": 131584,
    "temperature": 1.0,
    "top_p": 1.0,
    "reasoning_effort": "high"
  }'

https://huggingface.co/nota-ai/Solar-Open2-250B-Nota-NVFP4#citationCitation

@inproceedings{park2026dreammoe,
  title     = {{DREAM-MoE}: Downstream Routing Error-Aware Margin-Preserving Quantization for Mixture-of-Experts Large Language Models},
  author    = {Park, Hancheol and Lee, Geonho and Kim, Tae-Ho},
  booktitle = {ICML 2026 Workshop on Resource-Adaptive Foundation Model Inference (AdaptFM)},
  year      = {2026},
  url       = {https://openreview.net/forum?id=Wyhqwjl51A},
}

@inproceedings{lee2026sramoe,
  title     = {{SRA-MoE}: Output-Aware Selective Router Alignment for MoE Quantization},
  author    = {Lee, Geonho and Park, Hancheol and Lee, Seunghyun and Choi, Jungwook and Kim, Tae-Ho},
  booktitle = {ICML 2026 Workshop on Resource-Adaptive Foundation Model Inference (AdaptFM)},
  year      = {2026},
  url       = {https://openreview.net/forum?id=H0NoX02erJ},
}