BrowseComp — Generic 'multi-agent' descriptor, not confirmed identical to named 'Agent Swarm' scaffold
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Claude Mythos 5 | Anthropic | 93.3 | 2 | 2026-06-12 |
| 2 | Claude Opus 4.8 | Anthropic | 88.5 | 1 | 2026-06-10 |
| 3 | Claudes | Anthropic | 86.6 | 1 | 2026-08-12 |
| 4 | Claude Opus 4.6 | Anthropic | 86.6 | 1 | 2026-06-03 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.