Claude Opus 5
- Made by
- Anthropic
- Released
- Announced in
- Introducing Claude Opus 5anthropic.com
- Event
- Claude Opus 5
- API record
- /api/v1/models/claude-opus-5
Benchmark results
| Benchmark | Score | Source check |
|---|---|---|
| Humanity's Last ExamNo tools | 56.3% | CheckedA person checked this number against its source on 16 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| reported later, in Claude Opus 5.5 System Card | 56.6% | CheckedA person checked this number against its source on 16 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| Humanity's Last ExamWith tools | 64.7% | CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.CorroboratedThe same number is printed in 1 other lab document: 1 |
| reported later, in Introducing Claude Fable 5.1 and Claude Mythos 5.1 | 63.6% | CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| SWE-bench Proadaptive thinking, max effort; average of 5 trials · The source doesn't say whether this is the public or the commercial set | 79.2% | CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| BrowseCompcompaction at 200K tokens | 90.8% | CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| ARC-AGI-2max effort; semi-private set; as reported by the ARC Prize Foundation | 90.4% | CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| Terminal-Bench 4.0reported later, in Introducing Claude Fable 5.1 and Claude Mythos 5.1 | 52.3% | CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| AutomationBenchmax effort; Zapier leaderboard, private held-out set | 26.0% | CorroboratedThe same number is printed in 1 other lab document: 1 |
| medium effort; Zapier leaderboard, private held-out set · in Claude Opus 5 System Card | 24% | CheckedA person checked this number against its source on 20 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| Zapier leaderboard, private held-out set; max effort · reported later, in Introducing Claude Fable 5.1 and Claude Mythos 5.1 | 26.9% | |
| OSWorld 2.1Full set, partial score · max effort; mean of 5 runs; September 10, 2026 task files; the system card labels it OSWorld 2.0 · reported later, in Introducing Claude Opus 5.5 | 74.0% | CorroboratedThe same number is printed in 1 other lab document: 1 |
| OSWorld 2.0Offline set, partial score · task release not stated | 70.6% | CheckedA person checked this number against its source on 18 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| partial score; August 2026 task release, not comparable to earlier OSWorld 2.0 scores · reported later, in Introducing Claude Fable 5.1 and Claude Mythos 5.1 | 75.4% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.CorroboratedThe same number is printed in 1 other lab document: 1 |
| FrontierCode 1.1Main set · run and scored by Cognition; medium effort, the best of five effort levels; mean of 5 runs | 53.4% | |
| run by Cognition in Claude Code · reported later, in Introducing Claude Opus 5.5 | 48.0% | CorroboratedThe same number is printed in 1 other lab document: 1 |
| FrontierCode 1.1Extended set · run and scored by Cognition; medium effort, the best of five effort levels; mean of 5 runs | 63.6% | CorroboratedThe same number is printed in 1 other lab document: 1 |
| DeepSWE v1.1average of 5 trials | 68.8% | CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.CorroboratedThe same number is printed in 1 other lab document: 1 |
| ARC-AGI-3 | 30.2% | CheckedA person checked this number against its source on 18 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
Sources
Reported by Anthropic
- Claude Opus 5 System Cardanthropic.com · 24 Jul 2026
Our capture of pages 152-153 
Our capture of page 152 - Introducing Claude Opus 5anthropic.com · 24 Jul 2026
Original image - Introducing Claude Fable 5.1 and Claude Mythos 5.1anthropic.com · 1 Sep 2026
Our capture of the source - Introducing Claude Opus 5.5anthropic.com · 22 Sep 2026
Our capture of the source - Claude Opus 5.5 System Cardanthropic.com · 22 Sep 2026
Our capture of page 174
