How to read this page
LiveBench reports category scores on objective tasks. Arena reports blind human-preference Elo. SWE-bench Verified reports the rate at which an agent configuration resolves real repository issues.
Independent benchmark snapshots
Compare evidence by real work scenario
Scores remain separated by source and metric. They are evidence to inspect, not a universal model ranking.
Latest available source results
Snapshots refresh daily. A failed source never overwrites the last valid data. Exact model versions, test setup, and original leaderboard links are retained for each result.
Retrieved: Jul 19, 2026, 8:44 PM
Arena · code
Open original leaderboardArena has no official public API. Data is read from a daily structured mirror and always links back to the official leaderboard.
| Rank | Exact model | Score | Metric | Samples / votes |
|---|---|---|---|---|
| 1 | kimi-k3 | 1679 | Elo (Elo) | 1757 |
| 2 | claude-fable-5 | 1631 | Elo (Elo) | 2505 |
| 3 | gpt-5.6-sol-xhigh (codex-harness) | 1618 | Elo (Elo) | 2542 |
| 4 | glm-5.2 (max) | 1587 | Elo (Elo) | 4722 |
| 5 | claude-opus-4-8-thinking | 1562 | Elo (Elo) | 7309 |
| 6 | grok-4.5 | 1558 | Elo (Elo) | 2214 |
| 7 | claude-opus-4-7-thinking | 1558 | Elo (Elo) | 10534 |
| 8 | claude-opus-4-7 | 1555 | Elo (Elo) | 9976 |
| 9 | claude-opus-4-6-thinking | 1542 | Elo (Elo) | 12919 |
| 10 | claude-sonnet-5-high | 1542 | Elo (Elo) | 2959 |
Arena · search
Open original leaderboardArena has no official public API. Data is read from a daily structured mirror and always links back to the official leaderboard.
| Rank | Exact model | Score | Metric | Samples / votes |
|---|---|---|---|---|
| 1 | claude-opus-4-6-search | 1255 | Elo (Elo) | 105728 |
| 2 | gpt-5.5-search | 1239 | Elo (Elo) | 61253 |
| 3 | claude-opus-4-7 | 1235 | Elo (Elo) | 62150 |
| 4 | claude-fable-5 | 1230 | Elo (Elo) | 9895 |
| 5 | ernie-5.1 | 1227 | Elo (Elo) | 3803 |
| 6 | claude-sonnet-4-6-search | 1222 | Elo (Elo) | 105370 |
| 7 | gemini-3.1-pro-grounding | 1211 | Elo (Elo) | 83445 |
| 8 | claude-opus-4-8 | 1207 | Elo (Elo) | 42248 |
| 9 | gemini-3-pro-grounding | 1207 | Elo (Elo) | 37277 |
| 10 | gpt-5.2-search | 1206 | Elo (Elo) | 52708 |
Important limits
Scores from different benchmarks cannot be added together. The everyday-life view uses a preference proxy because no general authoritative benchmark covers that scenario.