Unverified50% confidenceBenchmarkExact time
Under the same 1000-agent configuration, Opus 5.5 ran for 30 hours and completed only 26% of the formal proof task.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
11/12/2026
First Seen
Valid until: 11/26/2026
Sources
Claude Fable 5.2 Hands-On: 1,000 Agents Prove Fermat's Last Theorem in 30 Hours
bilibili最佳拍档大飞11/12/2026
Related Entities
Related Claims
UnverifiedA team of 1000 Fable 5.2 agents completed a full Lean formalization of Fermat's Last Theorem in 30 hours; under the same configuration, Opus 5.5 completed only 26% in 18 hours and made almost no further progress by 30 hours.77% similarUnverifiedIn preliminary multi-agent experiments given a full 8 hours, the programming benchmark peaked at only 30.04 (versus 20.39 for single-agent), and the science-category agent benchmark reached only 31.2, lagging behind a competitor's 51.8.72% similarUnverifiedIn real tests of three related coding tasks, using the $20 Claude Code plan, Opus consumed 16% of the 5-hour quota while Haiku consumed only about 1%, and all six tasks passed testing and browser checks.69% similarUnverifiedThe training scale is 4096 games in parallel, about 330,000 agent steps per second end-to-end, completing about 19 billion agent steps in about 30 hours on a single Lambda GH200.69% similarUnverifiedBreaking tasks down into small steps of under 30 minutes with clear completion checkpoints is essentially about reducing extraneous cognitive load.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1163369API
curl https://kongchang.com/api/v1/knowledge/claims/1163369MCP
get_claim(id=1163369)