nvidia / text-generation updated 9 months ago

Qwen2.5-VL-7B-Instruct-NVFP4

The NVIDIA Qwen2.5-VL-7B-Instruct-FP4 model is the quantized version of Alibaba's Qwen2.5-VL-7B-Instruct model, which is an auto-regressive language model that uses an optimized transformer architecture. For more information, please check here. The NVIDIA Qwen2.5-VL-7B-Instruct-FP4 model is quantized with TensorRT Model Optimizer.

Params
5.0 B
Context
128,000
Downloads 30d
341 K
Likes
16
Licence: other Not gated SAFETENSORS View on Hugging Face ↗

Download history

tracking started — chart appears after 7 days of snapshots (3 recorded)
341 K downloads in the last 30 days

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

FileQuantSizeEst. VRAMVerdict on RTX 4090 · 24 GB
model.safetensors u8 7.2 GB 9.2 GB ✅ Runs comfortably
model.safetensors (bf16, full) bf16 + 125K ctx 7.2 GB 20.2 GB ⚠️ Tight — reduce context
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Run it

copy-paste, exact tags checked against the Hub
~ · curl · api/v1
$ curl -s https://aimodelscomparison.com/api/v1/models/qwen2-5-vl-7b-instruct-nvfp4
{
  "hf_id": "nvidia/Qwen2.5-VL-7B-Instruct-NVFP4",
  "params_b": 5.03,
  "context_length": 128000,
  "license": { "id": "other", "commercial": "unknown" },
  "downloads_30d": 341268,
  "vram_estimates": [
    { "quant": "u8", "gb": 9.2 }
  ],
  "updated_at": "2026-09-18T01:00:30Z"
}
est. VRAM —on RTX 4090 · 24 GBJSON API →

Specifications

Architecture
Qwen2_5_VLForConditionalGeneration
Parameters
5.0 B
Tensor type
U8
Context length
128,000
Vocabulary
152,064
Layers / heads
28 / 28
Licence
other
First seen on the Hub
2025-09-10
Base model
Qwen2.5-VL-7B-Instruct
Training datasets
undisclosed
Added to our catalog
2026-09-18

Family

Base model and the most-downloaded derivatives in the catalog.

Compare with any text-generation model