Best local models for coding

Open-weight models trained or tuned for code, with the exact size of every quantization so you can pick one that fits the machine you write on.

From the file· live filter over real dataFrom the file· 40 models

How this is ranked

Filtered to models whose publisher identifies them as code models, ranked by downloads. We do NOT rank on coding benchmarks: we have not measured them, and reproducing someone else's leaderboard under our own name would be the kind of borrowed authority this site exists to avoid.

Best local models for coding

#ModelParamsSmallest quantSmallest quant
1Qwen3-Coder-30B-A3B-InstructMoE30.5B7.46 GiB7.46 GiB
2KAT-Coder-V2.5-DevMoE34.7B9.11 GiB9.11 GiB
3Qwopus3.6-27B-Coder27.8B8.89 GiB8.89 GiB
4Qwen3-Coder-NextMoE79.7B15.44 GiB15.44 GiB
5Kimi-K2.7-CodeMoE1059B283.04 GiB283.04 GiB
6Qwen2.5-Coder-7B-Instruct7.6B2.59 GiB2.59 GiB
7Qwen2.5-Coder-32B-Instruct32.8B8.41 GiB8.41 GiB
8Qwen2.5-Coder-14B-Instruct14.8B4.38 GiB4.38 GiB
9Qwen2.5-Coder-3B-Instruct3.1B1.06 GiB1.06 GiB
10Qwen2.5-Coder-1.5B-Instruct1.5B0.56 GiB0.56 GiB
11Devstral-Small-2-24B-Instruct-251224.0B5.18 GiB5.18 GiB
12Yi-Coder-9B-Chat8.8B1.88 GiB1.88 GiB
13Yi-Coder-1.5B-Chat1.5B0.46 GiB0.46 GiB
14Huihui-Qwen3-Coder-Next-abliteratedMoE79.7B15.07 GiB15.07 GiB
15North-Mini-Code-1.0MoE30.5B7.93 GiB7.93 GiB
16Codestral-22B-v0.122.2B4.50 GiB4.50 GiB
17deepseek-coder-6.7b-instruct6.7B2.36 GiB2.36 GiB
18DeepSeek-Coder-V2-Lite-InstructMoE15.7B4.65 GiB4.65 GiB
19gemma-4-12B-coder-fable5-composer2.5-v1-uncensored-heretic12.0B2.78 GiB2.78 GiB
20Qwen3-53B-A3B-2507-THINKING-TOTAL-RECALL-v2-MASTER-CODERMoE53.0B10.22 GiB10.22 GiB
21Qwen2.5-Coder-14B-Instruct-abliterated14.8B4.38 GiB4.38 GiB
22Qwen2.5-Coder-0.5B-Instruct494M0.32 GiB0.32 GiB
23Jan-code-4b4.4B1.56 GiB1.56 GiB
24Qwen2.5-Coder-0.5B494M0.32 GiB0.32 GiB
25Huihui-Qwen3-Coder-30B-A3B-Instruct-abliteratedMoE30.5B5.98 GiB5.98 GiB
26Devstral-Small-250723.6B5.18 GiB5.18 GiB
27Qwen2.5-Coder-32B32.8B6.77 GiB6.77 GiB
28HopCoder-Mini-35B-A3B-VL36MoE35.1B8.82 GiB8.82 GiB
29gemma-4-12B-coder-fable5-composer2.5-v1-abliterated12.0B6.77 GiB6.77 GiB
30gemma-4-A4B-98e-v7-coder-itMoE20.5B8.22 GiB8.22 GiB
31deepseek-coder-33b-instruct33.3B11.51 GiB11.51 GiB
32Qwen2.5-Coder-3B3.1B1.06 GiB1.06 GiB
33Qwen3-Coder-Next-REAMMoE60.3B11.67 GiB11.67 GiB
34gemma-4-A4B-98e-v7-coderx-itMoE20.5B8.22 GiB8.22 GiB
35flux2-klein-9b-uncensored-text-encoder8.2B3.06 GiB3.06 GiB
36gemma-4-12B-coder-fable5-composer2.5-v112.0B2.78 GiB2.78 GiB
37deepseek-coder-1.3b-instruct1.3B0.40 GiB0.40 GiB
38OmniCoder-9B9.4B3.36 GiB3.36 GiB
39Qwable-5-27B-Coder27.8B6.98 GiB6.98 GiB
40Qwen2.5-Coder-32B-Instruct-abliterated32.8B9.27 GiB9.27 GiB
Spec sheetFrom the fileFrom the filewhat these mean