Claude Mythos Preview
- Made by
- Anthropic
- Released
- Announced in
- Claude Mythos Preview's cybersecurity capabilitiesanthropic.com
- Event
- Claude Mythos Preview
- API record
- /api/v1/models/claude-mythos-preview
Benchmark results
| Benchmark | Score | Source check |
|---|---|---|
| Humanity's Last ExamNo tools · reported later, in Introducing Claude Opus 4.7 | 56.8% | 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. |
| Humanity's Last ExamWith tools · reported later, in Introducing Claude Opus 4.7 | 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 |
| SWE-bench Verifiedreported later, in Introducing Claude Opus 4.7 | 93.9% | 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 |
| SWE-bench Proreported later, in Introducing Claude Opus 4.7 · The source doesn't say whether this is the public or the commercial set | 77.8% | 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. |
| GPQA Diamondreported later, in Introducing Claude Opus 4.7 | 94.6% | 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.CorroboratedThe same number is printed in 1 other lab document: 1 |
| BrowseCompadaptive thinking, max effort; 3M-token limit, compaction at 200K tokens | 86.9% | CorroboratedThe same number is printed in 1 other lab document: 1 |
| reported later, in Introducing Claude Opus 4.7 | 86.9% | 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. |
| reported later, in Claude Fable 5 and Claude Mythos 5 System Card | 87.9% | 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. |
Sources
Reported by Anthropic
- Claude Mythos Preview System Cardanthropic.com · 7 Apr 2026
- Introducing Claude Opus 4.7anthropic.com · 16 Apr 2026
Original image - Claude Fable 5 and Claude Mythos 5 System Cardanthropic.com · 9 Jun 2026
Our capture of page 251
