llamaperf

Best local LLM for coding

The models people report running for coding on their own hardware, ranked from community reports. Pick your memory size first: a model that is the best choice on a 24 GB card is often not the best on 12 GB.

Ranked from 158 coding reports on llamaperf.

By VRAM

All memory sizes

#ModelFastest runMedian t/sFastest t/sReports
1Qwen3.8Alibaba27B · NVFP4 on RTX 50904020170
2Qwen3.6Alibaba35B · 3B active · ninfer quant on RTX 30904817145
3DeepSeek V4 FlashDeepSeek284B · 13B active1618016
4MuseMeta30B · UD-Q5_K_M on RTX 5090852537
5Gemma 4Google DeepMind26B · 4B active on RTX 40901111384
6Ornith1.59B · Q6_K30362
7Qwen2.5Alibaba27B · Q6_K_XL on RTX 309049702
8GLM-5.2Zhipu AINVFP4 on DGX Spark10152
9Nex-N2.5-miniMLX-4bit on M5 Max 128GB1341341
10Qwen3Alibaba8B · Q4 on M2 Max 96GB43431
11Agents-A1InternScience35B · 3B active · ternary22221
12GLM-5.3Zhipu AIQ4 on M3 Ultra 256GB37371
13Qwen3.5Alibaba9B · Q4 on D700 12GB11111
14Ling-3.0Ant GroupINT4 on DGX Spark41411
15Qwen3-Coder-NextUD-Q6_K_XL on AMD Strix Halo 128GB37371
16DeepSeek V4.1 FlashoQ4e on M3 Ultra 512GB20201
17KAT-Coder V2.5-DevKwaipilotRTX 3060 12GB14141
18Mimo 2.5XiaomiRTX Pro 6000 Blackwell--1

Frequently asked

What is the best local LLM for coding?

It depends on how much memory you have, which is why this page is split by VRAM. Each tier ranks the models people report using for coding on that much memory, with the quant and GPU of the fastest run, so you can pick from setups like yours rather than from a single global list.

How are these rankings built?

From community performance reports tagged as coding use. Models are scored on how many reports they have, the fastest generation speed within the tier, how recent the reports are, and how complete the best report is. Nothing is sponsored.

Does a coding model need more VRAM than a chat model?

Not for the weights, but coding sessions carry long contexts: whole files, diffs and tool output. The KV cache for that context has to fit next to the weights, so leave more headroom than a chat use would need, or pick a quant one rung smaller.