All AI labs
One page per lab, with the score breakdown, model ranks and every sourced figure.
Data as of the week of 26 Oct 2026 (updated 26 Oct 2026, 08:09, Bucharest time). Every figure links to its source; see the methodology.
| Rank | Lab | HQ | Parent | Score | Best Arena model |
|---|---|---|---|---|---|
| 1 | Anthropic | San Francisco, US | n/a | 97.2 (5/5) | Claude Opus 5.5 (High) |
| 2 | Google DeepMind | London / Mountain View | Alphabet (public) | 86.9 (3/5) | Gemini 4 Argon (High) |
| 3 | OpenAI | San Francisco, US | n/a | 68.6 (5/5) | GPT-5.6 Sol (xHigh) |
| 4 | Meta Superintelligence Labs | Menlo Park, US | Meta Platforms (public) | 67.5 (3/5) | Muse Wonder (Max) |
| 5 | xAI (SpaceXAI) | Palo Alto / Memphis, US | SpaceX (acquired Feb 2026) | 58.3 (5/5) | Grok 4.20 Beta 1 |
| 6 | Alibaba (Qwen) | Hangzhou, CN | Alibaba Group (public) | 53.6 (4/5) | Qwen3.8-Max |
| 7 | Moonshot AI (Kimi) | Beijing, CN | n/a | 51.8 (5/5) | Kimi K3 (Max) |
| 8 | Z.ai (Zhipu) | Beijing, CN | n/a | 48.8 (3/5) | GLM-5.3 (Max) |
| 9 | DeepSeek | Hangzhou, CN | n/a | 45.3 (5/5) | DeepSeek-V4.1-Flash-20260910 (Max) |
| 10 | Mistral AI | Paris, FR | n/a | 27.7 (5/5) | Mistral Large 4 |
FAQ
Which AI labs does the scoreboard track?
10 labs: Anthropic, Google DeepMind, OpenAI, Meta Superintelligence Labs, xAI (SpaceXAI), Alibaba (Qwen), Moonshot AI (Kimi), Z.ai (Zhipu), DeepSeek, Mistral AI.
Where do the numbers come from?
Arena and Artificial Analysis for model rankings, Polymarket for odds, and public reporting for revenue, valuation and compute. Every figure on a lab page links to its source; see the methodology.