Vibe Code Bench
Results
26 of 41 tracked models have a result.
- 1
Anthropic · Max effort
92.4% - 2Gemini 4 ArgonLimited
Google DeepMind · High effort
91.9% - 3
Anthropic · Max effort
90.4% - 4
Anthropic · Max effort
90.3% - 5
Anthropic · Max effort
90.3% - 6
OpenAI · Max effort
89.6% - 7
OpenAI · Max effort
88.9% - 8
SpaceXAI · xHigh effort
86.2% - 9
Meta · Max effort
85.9% - 10MiMo-V2.6-ProOpen
Xiaomi
85.2% - 11Kimi K3Open
Moonshot AI · Max effort
85.0% - 12
DeepSeek · High effort
84.7% - 13DeepSeek V4 ProOpen
DeepSeek · Max effort
82.3% - 14
OpenAI · Max effort
81.7% - 15MiMo-V2.6-FlashOpen
Xiaomi
79.0% - 16
Google DeepMind · High effort
78.7% - 17
Mistral AI · High effort
78.4% - 18GLM-5.3Open
Z.ai · Max effort
78.1% - 19Hy4 previewOpen
Tencent
77.5% - 20
OpenAI · Max effort
74.6% - 21
StepFun
70.4% - 22
Alibaba
64.7% - 23MiniMax-M3Open
MiniMax
47.6% - 24
Google DeepMind · High effort
32.0% - 25InklingOpen
Thinking Machines Lab
19.2% - 26Inkling-SmallOpen
Thinking Machines Lab
19.1%
15 tracked models without a result
Command A+, Gemini 3.5 Flash-Lite, Gemma 4 31B, GLM-5.3-Flash, gpt-oss-120b, Hy3, Kimi K2.7 Code, MAI-Thinking-1, Mistral Medium 3.5, Mistral Small 4, Muse Glimmer 30B, Nemotron 3 Ultra, Nova 2 Lite, Qwen3.8-27B, Qwen3.8-Flash.
Missing means the evaluator has not published a run, not that the model would score poorly.