Updated 2026-09-20 · ranked by real download data

Best open coding models

Ranked nightly from our download snapshots of the Hugging Face catalog. Every entry shows its licence and what hardware it realistically needs.

#ModelParamsContextCommercial use30dMin VRAM
01 Qwen3-Coder-30B-A3B-Instruct-GGUF +4 variants unsloth · text-generation ✓ apache-2.0 12.7 M from 12.9 GB
02 Qwen2.5-Coder-7B-Instruct +6 variants Qwen · text-generation 7.6 B 33 K ✓ apache-2.0 2.6 M from 18.4 GB
03 Qwen2.5-Coder-14B-Instruct +1 variant Qwen · text-generation 14.8 B 33 K ✓ apache-2.0 1.8 M from 35.2 GB
04 Qwen2.5-Coder-32B-Instruct-AWQ +1 variant Qwen · text-generation 32.8 B 33 K ✓ apache-2.0 1.5 M from 26.7 GB
05 Qwen3-Coder-Next-FP8 +2 variants Qwen · text-generation 79.7 B 262 K ✓ apache-2.0 1.4 M from 100.9 GB
06 DeepSeek-Coder-V2-Lite-Instruct +1 variant deepseek-ai · text-generation 15.7 B 164 K other 908 K from 37.4 GB
07 Qwen3-Coder-480B-A35B-Instruct-FP8 Qwen · text-generation 480.2 B 262 K ✓ apache-2.0 804 K from 602.9 GB
08 deepseek-coder-7b-instruct-v1.5 deepseek-ai · text-generation 6.9 B 4 K other 722 K from 16.7 GB
09 Qwen2.5-Coder-7B Qwen · text-generation 7.6 B 33 K ✓ apache-2.0 719 K from 18.4 GB
10 Qwen2.5-Coder-3B-Instruct Qwen · text-generation 3.1 B 33 K other 599 K from 7.8 GB
11 Qwen2.5-Coder-1.5B-Instruct Qwen · text-generation 1.5 B 33 K ✓ apache-2.0 570 K from 4.1 GB
12 Qwen3-Coder-Next-NVFP4-GB10 ucbye · text-generation 262 K ✓ apache-2.0 515 K from 50.9 GB
13 gemma-4-12B-coder-fable5-composer2.5-v1-GGUF yuxinlu1 · text-generation ✓ apache-2.0 500 K from 5.8 GB
14 Kwaipilot_KAT-Coder-V2.5-Dev-GGUF bartowski · text-generation ✓ apache-2.0 396 K from 11.3 GB
15 Qwen2.5-Coder-1.5B Qwen · text-generation 1.5 B 33 K ✓ apache-2.0 349 K from 4.1 GB
16 CodeLlama-7b-hf codellama · text-generation 6.7 B 16 K ⚠ llama2 325 K from 16.3 GB
17 xflux_text_encoders XLabs-AI · text-generation 4.8 B ✓ apache-2.0 320 K from 11.7 GB
18 deepseek-coder-6.7b-instruct deepseek-ai · text-generation 6.7 B 16 K other 306 K from 16.3 GB
19 tiny_starcoder_py bigcode · text-generation 160 M bigcode-openrail-m 281 K from 1.2 GB
20 Yi-Coder-9B-Chat-GGUF MaziyarPanahi · text-generation unknown 228 K from 2.7 GB
Membership and ranking refresh nightly after the snapshot run. VRAM is an estimate for the smallest quantization at 8K context — methodology.