Monitor AI model performance over time. See how GPT, Claude, Gemini, and others compete as the landscape evolves.
Anthropic’s new Mythos-class public model now leads both boards. Claude Fable 5 is #1 on text with 1510 Elo and #1 on code with a massive 1665 Elo.
Read our analysis→LMArena Elo ratings by release date
← Swipe chart to explore · Tap any model for details →
Which company held the #1 spot over time
All tracked models ranked by Elo score
| Rank | Model | Score | Details |
|---|---|---|---|
| 1 | Claude Fable 5 Anthropic | 1508 | |
| 2 | Claude Opus 4.6 Thinking Anthropic | 1504 | |
| 3 | Claude Opus 4.7 Thinking Anthropic | 1502 | |
| 4 | Claude Opus 4.6 Anthropic | 1499 | |
| 5 | Claude Opus 4.7 Anthropic | 1493 | |
| 6 | Muse Spark Meta | 1487 | |
| 7 | Gemini 3.1 Pro Preview Google | 1486 | |
| 8 | Gemini 3 Pro Google | 1486 | |
| 9 | Claude Opus 4.8 Thinking Anthropic | 1483 | |
| 10 | GPT-5.5 High OpenAI | 1481 | |
| 11 | GPT-5.4 High OpenAI | 1478 | |
| 12 | Claude Opus 4.8 Anthropic | 1478 | |
| 13 | Gemini 3.5 Flash Google | 1476 | |
| 14 | GPT-5.2 Chat Latest OpenAI | 1475 | |
| 15 | GLM-5.1 Z.ai | 1475 | |
| 16 | GPT-5.5 OpenAI | 1475 | |
| 17 | Qwen3.7 Max Preview Alibaba | 1475 | |
| 18 | Grok 4.20 Beta 0309 Reasoning xAI | 1474 | |
| 19 | Grok 4.20 Beta1 xAI | 1474 | |
| 20 | Gemini 3 Flash Google | 1473 |
Trusted Data Sources
As of June 11, 2026, Claude Fable 5 holds the #1 position on text with 1510 Elo, followed by Claude Opus 4.6 Thinking (1504) and Claude Opus 4.7 Thinking (1502). On code, Claude Fable 5 leads with 1665 Elo, ahead of Claude Opus 4.7 Thinking at 1566 and Claude Opus 4.8 Thinking at 1559.
An Elo score is a rating system where models compete head-to-head. Users compare responses without knowing which model produced them, and scores adjust based on wins and losses.
Leadership has changed multiple times. OpenAI led with GPT-4 through most of 2024. Google took the lead with Gemini 3 Pro in January 2026. Anthropic has led since February 2026, and Claude Fable 5 now leads both the text and code leaderboards after its June 2026 debut.
We capture daily snapshots to track how rankings evolve over time. This reveals score drift as new models enter the arena.