Local-first ranking

The Local AI Leaderboard

Artificial Analysis ranks raw frontier intelligence - and almost all of it is proprietary, cloud-only, and metered by the token. We flip it: this board ranks the best models you can actually self-host, defaulting to open weights and surfacing the VRAM and params you need to run them. Same Intelligence / Coding / Math / Agentic indexes - reframed for $0/mo local instead of pay-per-call cloud.

Best you can run locally

Kimi K2.6 leads the open field at 53.9 Intelligence. Buy the GPU once, then $0/mo forever. No tokens, no rate limits, your data stays home.

Frontier, but cloud-only

Claude Opus 4.8 (max) tops the proprietary board at 61.4 - but you can't download it. It's metered per token, can change or vanish overnight, and every prompt leaves your machine. Use the toggle below to compare it against what you can self-host.

open / runs locally · $0/mo proprietary - not local
1
Kimi K2.6
1000B
53.9
2
MiMo-V2.5-Pro
1023.2B~593 GB
53.8
3
DeepSeek V4 Pro
1600B
51.5
4
GLM-5.1
744B
51.4
5
MiniMax-M2.7
228.7B~133 GB
49.6
6
MiMo-V2.5
310.8B~180 GB
49.0
7
Nemotron 3 Ultra
253.4B~147 GB
47.7
8
DeepSeek V4 Flash
284B
46.5
9
Qwen3.6 27B
28B~16 GB
45.8
10
Qwen3.5 397B A17B
403.4B~234 GB
45.0
11
Qwen3.6 35B A3B
36B~21 GB
43.5
12
Qwen3.5 122B A10B
125.1B~73 GB
41.6
13
Gemma 4 31B
32.7B~19 GB
39.2
14
NVIDIA Nemotron 3 Super
67.2B~39 GB
36.0
15
gpt-oss-120b
120.4B~70 GB
33.3
16
Qwen3.5 9B
9.7B~6 GB
32.4
17
Gemma 4 26B A4B
26.5B~15 GB
31.2
18
Qwen3.5 35B A3B
36B~21 GB
30.7
19
EXAONE 4.5 33B
34.4B~20 GB
30.2
20
Nemotron Cascade 2 30B A3B
31.6B~18 GB
28.4
21
Qwen3-Coder-Next
79.7B~46 GB
28.3
22
Qwen3.5 4B
4.7B~3 GB
27.1
23
Qwen3-Next-80B-A3B-Instruct
81.3B~47 GB
26.7
24
gpt-oss-20b
21.5B~12 GB
24.5
25
NVIDIA-Nemotron-Nano-9B-v2
8.9B~5 GB
14.8
26
Llama 3.3 70B
70B~40 GB
14.5
27
Phi-4
14B~8.5 GB
10.4
28
Phi-4 Mini
3B~2.5 GB
8.4
29
LFM2.5-1.2B-Instruct
1.2B~1 GB
8.0
30
gemma-3-270m
0.3B
7.7
30 open models you can run locally · sorted by Intelligence IndexData: Artificial Analysis · updated 2026-06-07

Found a model that fits your GPU? Each one has a full page with benchmarks, VRAM math, and recommended hardware.

Browse all local models