Road to AGI
Data as of

Model roster · 47 tracked systems

Every major contender, on one field.

This roster covers OpenAI, Anthropic, Google, xAI, Meta, Mistral, Microsoft, Moonshot, Alibaba/Qwen, Zhipu, DeepSeek, MiniMax, 01.AI, Baichuan, ByteDance and other tracked labs.

See roster and deduplication rules →

Comparable ECI records

Model leaderboard

“Open weights” follows the source’s access label. It does not automatically mean unrestricted, commercially usable, or fully open-source.

Showing 47 of 47 tracked models
ModelEcosystemLabAccessReleasedECI
Claude Fable 5 (max) Confident source confidence
WesternAnthropicAPI162.5 rank 1
GPT-5.5 Pro (xhigh) Likely source confidence
WesternOpenAIAPI161.7 rank 2
Claude Opus 5 (max) Confident source confidence
WesternAnthropicAPI161.6 rank 3
GPT-5.6 Sol (pro, max) Confident source confidence
WesternOpenAIAPI161.1 rank 4
GPT-5.6 Terra (max) Confident source confidence
WesternOpenAIAPI158.9 rank 5
Claude Opus 4.8 Confident source confidence
WesternAnthropicAPI157.7 rank 8
Kimi K3 (max) Speculative source confidence
ChineseMoonshot AIOpen weights · non-commercial157.3 rank 9
Gemini 3.7 Flash (high) Confident source confidence
WesternGoogle DeepMindAPI157.1 rank 10
Qwen3.8 Max (xhigh) Confident source confidence
ChineseAlibaba / QwenAPI156.4 rank 12
Grok 4.6 (xhigh) Confident source confidence
WesternxAIAPI156.2 rank 15
DeepSeek V4 Pro 0813 (max) Unverified source confidence
ChineseDeepSeekOpen weights · unrestricted155.2 rank 18
Muse Spark 1.2 (xhigh) Confident source confidence
WesternMeta AIAPI155.2 rank 19
Gemini 3.1 Pro Preview Likely source confidence
WesternGoogle DeepMindAPI154.6 rank 22
Gemini 3.5 Flash Confident source confidence
WesternGoogle DeepMindAPI154.4 rank 23
DeepSeek V4 Flash 0731 (max) Likely source confidence
ChineseDeepSeekOpen weights · unrestricted154.3 rank 24
Muse Spark 1.1 (high) Confident source confidence
WesternMeta AIAPI154.3 rank 25
Grok 4.5 (high) Likely source confidence
WesternxAIAPI153.9 rank 27
Qwen3.7 Max Confident source confidence
ChineseAlibaba / QwenAPI153.6 rank 28
Muse Spark Confident source confidence
WesternMeta AIAPI152.6 rank 31
Grok 4.20 Likely source confidence
WesternxAIAPI152.1 rank 33
GLM-5.2 Confident source confidence
ChineseZhipu AI / Z.aiOpen weights · unrestricted152.0 rank 34
Kimi K2.6 Confident source confidence
ChineseMoonshot AIOpen weights · unrestricted151.0 rank 36
GLM-5.1 Confident source confidence
ChineseZhipu AI / Z.aiOpen weights · unrestricted150.8 rank 37
Kimi K2.7 Code Confident source confidence
ChineseMoonshot AIOpen weights · unrestricted150.5 rank 38
Qwen 3.6 Plus (2026-04-02) Confident source confidence
ChineseAlibaba / QwenAPI149.4 rank 44
DeepSeek v4 Pro (max) Likely source confidence
ChineseDeepSeekOpen weights · unrestricted149.2 rank 47
MiniMax-M3 Confident source confidence
ChineseMiniMaxOpen weights · restricted use147.8 rank 52
MiniMax-M2.5 Likely source confidence
ChineseMiniMaxOpen weights · unrestricted146.8 rank 58
GLM-5 Likely source confidence
ChineseZhipu AI / Z.aiOpen weights · unrestricted146.2 rank 62
MiniMax-M2.7 Confident source confidence
ChineseMiniMaxOpen weights · non-commercial146.0 rank 64
Qwen3.6 35B-A3B Unverified source confidence
ChineseAlibaba / QwenOpen weights · unrestricted145.0 rank 69
Mistral Medium 3.5 Confident source confidence
WesternMistral AIOpen weights · restricted use142.5 rank 86
Gemma 4 31B IT Likely source confidence
WesternGoogle DeepMindOpen weights · restricted use142.4 rank 87
Mistral Medium 3 Unknown source confidence
WesternMistral AIAPI135.3 rank 108
Magistral Small 1.0 Confident source confidence
WesternMistral AIOpen weights · unrestricted134.2 rank 110
Llama 4 Maverick (FP8) Likely source confidence
WesternMeta AIOpen weights · restricted use133.0 rank 113
Phi-4 Confident source confidence
WesternMicrosoftOpen weights · unrestricted131.0 rank 115
phi-3-small 7.4B Confident source confidence
WesternMicrosoftOpen weights · unrestricted121.6 rank 142
phi-3-medium 14B Likely source confidence
WesternMicrosoftOpen weights · unrestricted121.0 rank 145
Yi-34B Confident source confidence
Chinese01.AIOpen weights · restricted use117.0 rank 156
Yi 6B Confident source confidence
Chinese01.AIOpen weights · restricted use104.2 rank 175
Baichuan2-13B Confident source confidence
ChineseBaichuanOpen weights · restricted use102.6 rank 176
Baichuan 2-7B Confident source confidence
ChineseBaichuanOpen weights · restricted use95.6 rank 183
Baichuan1-7B Confident source confidence
ChineseBaichuanOpen weights · non-commercial89.6 rank 190
GPT-2 (1.5B) Speculative source confidence
WesternOpenAIOpen weights · unrestricted0.0 rank 335
Yi-Lightning Confident source confidence
Chinese01.AIAPI0.0 rank 283
video-SALMONN 2+ Likely source confidence
ChineseByteDanceNot reported0.0 rank 249

Coverage gaps

Popular labs missing a comparable score

Xiaomi Chinese ecosystem

No valid scored ECI row in the current public dataset. The lab remains in the tracked scope and will appear when comparable data becomes available.

Baidu Chinese ecosystem

No valid scored ECI row in the current public dataset. The lab remains in the tracked scope and will appear when comparable data becomes available.

Tencent Chinese ecosystem

No valid scored ECI row in the current public dataset. The lab remains in the tracked scope and will appear when comparable data becomes available.