---
license: apache-2.0
library_name: transformers
pipeline_tag: image-text-to-text
base_model: Qwen/Qwen3.5-27B
base_model_relation: quantized
tags:
  - transformers
  - safetensors
  - qwen3_5
  - quantized
  - nvfp4
  - fp4
  - 4-bit
  - vllm
  - llm-compressor
  - image-text-to-text
  - conversational
datasets:
  - neuralmagic/calibration
---

# Qwen3.5-27B-NVFP4

This is a quantized version of [Qwen/Qwen3.5-27B](https://huggingface.co/Qwen/Qwen3.5-27B). This model accepts text and images as inputs and generates text as outputs. The weights and activations were quantized to FP4 using [llm-compressor](https://github.com/vllm-project/llm-compressor) with 512 calibration samples from [neuralmagic/calibration](https://huggingface.co/datasets/neuralmagic/calibration), reducing the model size from 51.8 GB to 18.4 GB (~2.8x reduction) while maintaining 99.1% average accuracy recovery.

---

## Inference

As of 2/27/2026, this model is supported in vLLM nightly. To serve the model:

```bash
vllm serve Kbenkhaled/Qwen3.5-27B-NVFP4 \
    --reasoning-parser qwen3 \
    --enable-prefix-caching
```

---

## Evaluation

Evaluated with [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness), 0-shot, thinking mode ON.

| Benchmark | Qwen3.5-27B | Qwen3.5-27B-NVFP4 (this model) | Recovery |
|---|---|---|---|
| GPQA Diamond | 80.30% | 79.29% | 98.7% |
| IFEval | 95.08% | 93.88% | 98.7% |
| MMLU-Redux | 93.90% | 94.32% | 100.4% |
| **Average** | **89.76%** | **89.16%** | **99.1%** |