liwaiwailiwaiwai labs

// leaderboard

AI Model Leaderboard

Ranked by benchmark score latency, cost-efficiency, accuracy, coding, and reasoning.

ExploreAI Tools

Best for All

Claude Sonnet 4 — Anthropic

🥇
#Model
🥇
Claude Sonnet 4Anthropic
92
95%94%1.1s$3.00
🥈
GPT-4.1OpenAI
91
94%92%980ms$2.00
🥉
Claude 3.7 SonnetAnthropic
90
93%93%1.2s$3.00
4
GPT-4oOpenAI
89
91%89%820ms$2.50
5
Gemini 2.5 ProGoogle
89
92%91%1.3s$3.50
6
DeepSeek V3DeepSeek
89
88%91%760ms$0.27
7
o4-miniOpenAI
88
88%90%1.1s$1.10
8
Gemini 2.5 FlashGoogle
87
85%84%550ms$0.35
9
DeepSeek R1DeepSeekfree
86
87%88%2.1sFree
10
Mistral LargeMistral
86
86%85%900ms$2.00
11
Qwen3 Coder 480BAlibabafree
85
82%90%1.2sFree
12
GPT-4o miniOpenAI
85
82%81%380ms$0.15
13
Claude 3.5 HaikuAnthropic
85
83%82%450ms$0.25
14
GPT OSS 120BOpenAIfree
84
84%83%1.5sFree
15
Llama 3.3 70BMetafree
82
80%79%870msFree
16
Nemotron Super 120BNVIDIAfree
82
80%81%1.1sFree
17
Gemma 3 27BGooglefree
81
78%76%900msFree
18
MiniMax M2.5MiniMaxfree
80
77%75%880msFree
19
o3OpenAI
80
97%96%4.2s$15.00
20
GPT OSS 20BOpenAIfree
79
75%74%680msFree
21
GLM 4.5 AirZ.aifree
ExploreAI Tools
78
73%72%520msFree
22
Gemma 3 12BGooglefree
77
72%70%500msFree
23
Nemotron Nano 30BNVIDIAfree
77
71%72%450msFree
24
Llama 3.2 3BMetafree
72
65%63%290msFree
Discover more
Robotics
Try CAD Software
Take Psychology Courses

Scores are composite benchmarks. Figures are representative and updated periodically.