Updated daily at 05:00 UTC
Ranking
Best AI models for Human preference
Which model people prefer in blind side-by-side votes (chat, web dev, vision, search, documents, agents, voices), reported as Elo with confidence intervals.
Ranked models
Based on 5 benchmarks
| Rank | Model | Company | Score | Coverage | Released |
|---|---|---|---|---|---|
| 1 | Grok 4.6 | SpaceXAI (xAI) | 64.8 | 3 of 5 | 12 Aug 2026 |
| 2 | Claude Fable 5.1 | Anthropic | 62.4 | 4 of 5 | 1 Sep 2026 |
| 3 | GPT-6 Astra | OpenAI | 61.5 | 4 of 5 | 3 Sep 2026 |
| 4 | Claude Opus 5 | Anthropic | 61.1 | 4 of 5 | 24 Jul 2026 |
| 5 | Kimi K2.6 | Moonshot AI | 60.3 | 3 of 5 | 20 Apr 2026 |
| 6 | Muse Spark 1.3 | Meta | 59.7 | 4 of 5 | 2 Sep 2026 |
| 7 | Claude Opus 4.5 | Anthropic | 59.6 | 3 of 5 | 24 Nov 2025 |
| 8 | Claude Opus 4.6 | Anthropic | 58.5 | 4 of 5 | 5 Feb 2026 |
| 9 | Gemini 3.7 Flash | Google DeepMind | 57.7 | 3 of 5 | 13 Aug 2026 |
| 10 | MiniMax-M3 | MiniMax | 57.4 | 3 of 5 | 1 Jun 2026 |
| 11 | Gemini 3.8 Flash | Google DeepMind | 57.0 | 3 of 5 | 2 Sep 2026 |
| 12 | Muse Spark 1.1 | Meta | 56.7 | 4 of 5 | 8 Apr 2026 |
| 13 | Kimi K2.5 | Moonshot AI | 55.8 | 3 of 5 | 27 Jan 2026 |
| 14 | Gemini 3.6 Flash | Google DeepMind | 55.3 | 4 of 5 | 21 Jul 2026 |
| 15 | Claude Sonnet 4.6 | Anthropic | 55.2 | 4 of 5 | 17 Feb 2026 |
| 16 | Claude Sonnet 5 | Anthropic | 55.0 | 4 of 5 | 30 Jun 2026 |
| 17 | Claude Fable 5 | Anthropic | 54.3 | 5 of 5 | 9 Jun 2026 |
| 17 | Gemini 3.5 Flash | Google DeepMind | 54.3 | 4 of 5 | 19 May 2026 |
| 19 | Claude Sonnet 4.5 | Anthropic | 54.2 | 3 of 5 | 29 Sep 2025 |
| 20 | Grok 4.20 | SpaceXAI (xAI) | 54.0 | 3 of 5 | Feb 2026 |
| 21 | Grok 4.5 | SpaceXAI (xAI) | 53.3 | 4 of 5 | 8 Jul 2026 |
| 22 | Gemini 3.1 Pro | Google DeepMind | 52.6 | 4 of 5 | 19 Feb 2026 |
| 23 | GPT-5.5 | OpenAI | 52.5 | 3 of 5 | 23 Apr 2026 |
| 24 | Gemini 3 Pro Retired | Google DeepMind | 52.4 | 4 of 5 | 18 Nov 2025 |
| 25 | Claude Opus 4.7 | Anthropic | 52.0 | 5 of 5 | 16 Apr 2026 |
| 26 | Gemini 3 Flash | Google DeepMind | 51.0 | 4 of 5 | 17 Dec 2025 |
| 27 | GPT-5.4 | OpenAI | 50.9 | 4 of 5 | 5 Mar 2026 |
| 28 | GPT-5.6 Terra | OpenAI | 50.7 | 3 of 5 | 9 Jul 2026 |
| 29 | Claude Opus 4.8 | Anthropic | 50.4 | 5 of 5 | 28 May 2026 |
| 30 | GPT-5.6 Luna | OpenAI | 49.1 | 3 of 5 | 9 Jul 2026 |
| 31 | Gemini 3.5 Flash-Lite | Google DeepMind | 48.3 | 3 of 5 | 21 Jul 2026 |
| 32 | Claude Haiku 4.5 | Anthropic | 48.2 | 3 of 5 | 15 Oct 2025 |
| 33 | Gemma 4 31B | Google DeepMind | 47.5 | 4 of 5 | 31 Mar 2026 |
| 33 | GPT-5.2 | OpenAI | 47.5 | 4 of 5 | 11 Dec 2025 |
| 35 | GPT-5.1 | OpenAI | 47.3 | 4 of 5 | 12 Nov 2025 |
| 36 | GPT-5.6 Sol | OpenAI | 47.2 | 4 of 5 | 9 Jul 2026 |
| 37 | GLM-5V-Turbo | Z.ai (Zhipu AI) | 45.5 | 4 of 5 | 2 Apr 2026 |
| 38 | GPT-5.4 mini | OpenAI | 45.4 | 3 of 5 | 17 Mar 2026 |
| 39 | GPT-5 Deprecated | OpenAI | 44.2 | 3 of 5 | 7 Aug 2025 |
| 40 | Gemma 4 26B A4B | Google DeepMind | 42.7 | 3 of 5 | 31 Mar 2026 |
| 41 | Inkling Small | Thinking Machines Lab | 42.2 | 3 of 5 | 30 Jul 2026 |
| 42 | Gemini 2.5 Pro | Google DeepMind | 41.8 | 4 of 5 | 25 Mar 2025 |
| 43 | Grok 4.3 | SpaceXAI (xAI) | 41.0 | 3 of 5 | 17 Apr 2026 |
| 44 | Gemini 3.1 Flash-Lite | Google DeepMind | 37.6 | 3 of 5 | 3 Mar 2026 |
Not ranked yet
These models have independent or third-party results on too few of the 5 benchmarks to rank.
