unsloth / text-generation updated 1 year ago

Qwen2.5-3B-Instruct-unsloth-bnb-4bit

Name: Qwen2.5-3B-Instruct-unsloth-bnb-4bit
Author: unsloth

See our collection for versions of Qwen2.5 including 4-bit formats. Unsloth's Dynamic 4-bit Quants is selectively quantized, greatly improving accuracy over standard 4-bit. Finetune LLMs 2-5x faster with 70% less memory via Unsloth!

Params

3.2 B

Context

32,768

Downloads 30d

449 K

Likes

Commercial use: allowed apache-2.0 Not gated SAFETENSORS 1 languages View on Hugging Face ↗

Download history

daily snapshots · 10 days

449 K424 K

Jul 28Jul 31Aug 3Aug 6

Can you run it?

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

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
model.safetensors	u8	2.4 GB	3.6 GB	✅ Runs comfortably
model.safetensors (bf16, full)	bf16 + 32K ctx	2.4 GB	5.0 GB	✅ Runs comfortably

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/qwen2-5-3b-instruct-unsloth-bnb-4bit
{
  "hf_id": "unsloth/Qwen2.5-3B-Instruct-unsloth-bnb-4bit",
  "params_b": 3.17,
  "context_length": 32768,
  "license": { "id": "apache-2.0", "commercial": "yes" },
  "downloads_30d": 449063,
  "vram_estimates": [
    { "quant": "u8", "gb": 3.6 }
  ],
  "updated_at": "2026-07-28T18:05:39Z"
}

est. VRAM —on RTX 4090 · 24 GBJSON API →

Specifications

Architecture: Qwen2ForCausalLM
Parameters: 3.2 B
Tensor type: U8
Context length: 32,768
Vocabulary: 151,936
Layers / heads: 36 / 16
Licence: apache-2.0
First seen on the Hub: 2025-02-06
Base model: Qwen2.5-3B-Instruct
Training datasets: undisclosed
Added to our catalog: 2026-07-28

Family

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

Base Qwen2.5-3B-Instruct Qwen 6.0 M

Compare with

Qwen3-8B 16.3 M Qwen2.5-1.5B-Instruct 14.0 M Qwen2.5-7B-Instruct 12.1 M Llama-3.2-1B-Instruct 10.4 M Qwen3-1.7B 7.8 M Qwen2.5-3B-Instruct 6.0 M

Appears in

Best open coding models Commercial-use text models Models that fit in 12 GB