Claude Opus 4.8: benchmarks and ranking
Claude Opus 4.8 (Anthropic) — overall rank: #43 / 153, overall score: 72.4 (90% interval 67.1–77.2), leaderboards: 7. Released: 2026-05-28 · Price (in/out): $5 / $25
Rank on each leaderboard
| Leaderboard | Rank | Published value |
|---|---|---|
| Artificial Analysis | #23 / 141 | 41.8 |
| LiveBench | #29 / 64 | 76.2 |
| Epoch AI | #23 / 230 | 159.9 |
| ARC Prize (ARC-AGI) | #28 / 95 | 72.1 |
| Scale AI SEAL | Not rated | — |
| Vals AI | Not rated | — |
| Terminal-Bench | #31 / 53 | 23.64 |
| EQ-Bench | #9 / 32 | 1267.8 |
| SimpleBench | #35 / 112 | 64.8 |
| BenchLM.ai | Not rated | — |
Domain scores
- Coding: 71.0
- Agents: 62.5
- Reasoning: 79.0
- Math: 77.8
- Knowledge: 83.1
- Writing: 70.0
- Instructions: 69.9
- Chat & EQ: 79.2
- Business: 71.3
- Japanese: 79.2
- Vision: 68.5
- Long context: 77.1