Frontier Log

AnthropicModel

Claude Opus 4.7

Made by
Anthropic
Released
Announced in
Introducing Claude Opus 4.7anthropic.com
Event
Claude Opus 4.7
API record
/api/v1/models/claude-opus-4-7

Benchmark results

BenchmarkScoreSource check
Humanity's Last ExamNo tools46.9%
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.
CorroboratedThe same number is printed in 4 other lab documents, the first 3 linked: 1 2 3
Humanity's Last ExamWith tools54.7%
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 4 other lab documents, the first 3 linked: 1 2 3
SWE-bench Verified87.6%
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.
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
SWE-bench Promemorization screens flag a subset of problems · The source doesn't say whether this is the public or the commercial set64.3%
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.
CorroboratedThe same number is printed in 6 other lab documents, the first 3 linked: 1 2 3
GPQA Diamond94.2%
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.
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
BrowseCompno thinking, max effort; 10M-token limit, compaction at 200K tokens79.3%
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.
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
re-run with a new blocklist, compaction at 200K tokens and adaptive thinking · reported later, in Claude Opus 4.8 System Card79.8%
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 thinking75.83%
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.
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
Terminal-Bench 2.0Terminus 2 harness, no thinking; averaged over 5 attempts per task69.4%
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.
CorroboratedThe same number is printed in 2 other lab documents: 1 2
OSWorld-Verified78.0%
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.
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
Sources

Reported by Anthropic

Events

Stories

All models