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

GLM-5.2 leads the open field at 51.1 Intelligence - ~437 GB VRAM to run. Buy the GPU once, then $0/mo forever. No tokens, no rate limits, your data stays home.

Frontier, but cloud-only

Claude Fable 5 (with fallback) tops the proprietary board at 59.9 - 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
GLM-5.2
753.4B~437 GB
51.1
2
MiniMax-M3
440.3B~255 GB
44.4
3
DeepSeek V4 Pro
1600B
44.3
4
MiMo-V2.5-Pro
1023.2B~593 GB
42.2
5
Kimi K2.7 Code
1058.6B~614 GB
41.9
6
Hy3
298.8B~173 GB
41.2
7
Nex-N2-Pro
396.8B~230 GB
41.0
8
Inkling
552.8B~321 GB
40.7
9
DeepSeek V4 Flash
284B
40.3
10
GLM-5.1
744B
40.2
11
MiniMax-M2.7
228.7B~133 GB
38.1
12
Nemotron 3 Ultra
253.4B~147 GB
37.8
13
MiMo-V2.5
310.8B~180 GB
37.2
14
Qwen3.6 27B
28B~16 GB
37.1
15
Kimi K2.6
1000B
34.6
16
Qwen3.5 397B A17B
403.4B~234 GB
33.7
17
Qwen3.5 122B A10B
125.1B~73 GB
32.3
18
Qwen3.6 35B A3B
36B~21 GB
31.6
19
Gemma 4 31B
32.7B~19 GB
29.4
20
Gemma 4 26B A4B
26.5B~15 GB
25.7
21
NVIDIA Nemotron 3 Super
67.2B~39 GB
25.4
22
Qwen3.5 35B A3B
36B~21 GB
24.0
23
gpt-oss-120b
120.4B~70 GB
23.8
24
EXAONE 4.5 33B
34.4B~20 GB
23.0
25
Gemma 4 12B
12B~7 GB
22.0
26
Qwen3.5 9B
9.7B~6 GB
21.4
27
Qwen3-Coder-Next
79.7B~46 GB
21.1
28
Qwen3.5 4B
4.7B~3 GB
20.1
29
North Mini Code
30.5B~18 GB
19.8
30
HyperNova 60B 2605
58.7B~34 GB
17.8
31
Nemotron Cascade 2 30B A3B
31.6B~18 GB
17.6
32
Qwen3-Next-80B-A3B-Instruct
81.3B~47 GB
16.7
33
gpt-oss-20b
21.5B~12 GB
14.9
34
MiniCPM5-1B
1.1B~1 GB
12.0
35
Llama 3.3 70B
70B~40 GB
9.4
36
NVIDIA-Nemotron-Nano-9B-v2
8.9B~5 GB
8.8
37
granite-4.1-8b
8.8B~5 GB
6.7
38
Phi-4 Mini
3B~2.5 GB
6.0
39
Phi-4
14B~8.5 GB
4.9
40
granite-4.1-3b
3.4B~2 GB
4.7
41
LFM2.5-1.2B-Instruct
1.2B~1 GB
2.7
42
gemma-3-270m
0.3B
2.4
42 open models you can run locally · sorted by Intelligence IndexData: Artificial Analysis · updated 2026-07-23

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

Browse all local models