Jackrong / text-generation updated 5 months ago

Qwen3.5-4B-Claude-4.6-Opus-Reasoning-Distilled-GGUF

🔥 Update (April 5): I’ve released the complete training notebook, codebase, and a comprehensive PDF guide to help beginners and enthusiasts understand and reproduce this model's fine-tuning process.

Params
Context
Downloads 30d
309 K
Likes
162
Commercial use: allowed apache-2.0 Not gated GGUF 3 languages View on Hugging Face ↗

Download history

daily snapshots · 34 days
▲ 96 K in the last 30 days (45.3%)
396 K179 K
Aug 18Aug 29Sep 9Sep 20

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
Qwen3.5-4B.Q2_K.gguf Q2_K 1.8 GB 2.5 GB ✅ Runs comfortably
Qwen3.5-4B.Q3_K_S.gguf Q3_K_S 2.1 GB 2.8 GB ✅ Runs comfortably
Qwen3.5-4B.Q3_K_M.gguf Q3_K_M 2.3 GB 3.0 GB ✅ Runs comfortably
Qwen3.5-4B.Q3_K_L.gguf Q3_K_L 2.4 GB 3.1 GB ✅ Runs comfortably
Qwen3.5-4B.Q4_K_S.gguf Q4_K_S 2.6 GB 3.3 GB ✅ Runs comfortably
Qwen3.5-4B.Q4_K_M.gguf Q4_K_M 2.7 GB 3.5 GB ✅ Runs comfortably
Qwen3.5-4B.Q5_K_S.gguf Q5_K_S 3.0 GB 3.8 GB ✅ Runs comfortably
Qwen3.5-4B.Q5_K_M.gguf Q5_K_M 3.1 GB 3.9 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
~ · ollama · Q4_K_M
$ ollama run qwen3-5-4b-claude-4-6-opus-reasoning-distilled-gguf

# pin the quantization explicitly
$ ollama run qwen3-5-4b-claude-4-6-opus-reasoning-distilled-gguf-q4_k_m
est. VRAM 3.5 GBon RTX 4090 · 24 GBJSON API →

Specifications

Architecture
Qwen3_5ForConditionalGeneration
Licence
apache-2.0
First seen on the Hub
2026-03-03
Base model
Qwen3.5-4B
Training datasets
nohurry/Opus-4.6-Reasoning-3000x-filtered, Jackrong/Qwen3.5-reasoning-700x
Added to our catalog
2026-08-18

Family

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

Compare with any text-generation model