Claude Mythos Preview
Claude / Anthropic
42scores
33benchmarks
$25 / $125 per 1M tokenscost in/out
Metadata
Claude Closed/API
Aliases: anthropic-claude-mythos-preview, anthropic/claude-mythos-preview, claude-mythos-preview, mythos-preview, Mythos Preview
| Benchmark | Category | Rank | Score | Sampled |
|---|---|---|---|---|
| BrowseComp | Agentic | 2 | 86.9% | 2026-04-16 |
| OSWorld-Verified | Agentic | 1 | 0.80 | 2026-05-06 |
| OSWorld-Verified | Agentic | 1 | 79.6% | 2026-04-16 |
| AAV Capsid Packaging Prediction | Biology | 1 | 0.8 | 2026-05-28 |
| BioPipelineBench Verified | Biology | 1 | 88.1% | 2026-05-28 |
| Black-box RNA Sequence Design | Biology | 1 | 11.2 | 2026-05-28 |
| DNA Synthesis Screening Evasion | Biology | 1 | 0.8 | 2026-05-28 |
| LABBench2 Clinical Trials | Biology | 1 | 86.3% | 2026-05-28 |
| LABBench2 Patent Questions | Biology | 2 | 64.3% | 2026-05-28 |
| Long-form Virology Task 1 | Biology | 1 | 0.8 | 2026-05-28 |
| Long-form Virology Task 2 | Biology | 1 | 0.9 | 2026-05-28 |
| ProteinGym Hard | Biology | 1 | 43.1% | 2026-05-28 |
| Protocol Troubleshooting (Anthropic Internal) | Biology | 1 | 69.6% | 2026-05-28 |
| scBench | Biology | 1 | 58.2% | 2026-05-28 |
| SpatialBench | Biology | 1 | 53.8% | 2026-05-28 |
| Structural Biology Open-Ended | Biology | 1 | 81.6% | 2026-05-28 |
| VCT Virology Capabilities Test | Biology | 1 | 0.6 | 2026-05-28 |
| Organic Chemistry (Anthropic Internal) | Chemistry | 1 | 86.5% | 2026-05-28 |
| Terminal-Bench 2.0 | Coding | 1 | 82% | 2026-04-16 |
| CyberGym | Cybersecurity | 1 | 83.1% | 2026-05-28 |
| CyberGym | Cybersecurity | 1 | 0.83 | 2026-05-06 |
| CyberGym | Cybersecurity | 1 | 83.1% | 2026-04-16 |
| ExploitBench v8-bench | Cybersecurity | 1 | 9.9 points | 2026-05-28 |
| ExploitBench v8-bench | Cybersecurity | 2 | 9.55 points | 2026-05-28 |
| ExploitBench v8-bench | Cybersecurity | 1 | 9.9 points | 2026-05-15 |
| ExploitBench v8-bench | Cybersecurity | 2 | 9.55 points | 2026-05-15 |
| Firefox 147 JS Exploitation | Cybersecurity | 1 | 70.8% | 2026-05-28 |
| BenchLM | General Knowledge | 1 | 99 | 2026-05-06 |
| Humanity's Last Exam | Intelligence | 1 | 64.7% | 2026-04-16 |
| Graphwalks BFS >128k | Long Context | 1 | 0.80 | 2026-05-06 |
| USAMO25 | Mathematics | 1 | 0.98 | 2026-05-06 |
| CharXiv-R | Multimodal | 1 | 0.93 | 2026-05-06 |
| CharXiv-R | Multimodal | 1 | 93.2% | 2026-04-16 |
| FigQA | Multimodal | 1 | 0.89 | 2026-05-06 |
| GPQA Diamond | Reasoning | 1 | 94.6% | 2026-04-16 |
| BioMysteryBench Human-Difficult | Science | 2 | 29.6% | 2026-05-28 |
| BioMysteryBench Human-Difficult | Science | 1 | 29.6% | 2026-04-29 |
| BioMysteryBench Human-Solvable | Science | 1 | 82.6% | 2026-05-28 |
| BioMysteryBench Human-Solvable | Science | 1 | 82.6% | 2026-04-29 |
| DeepSearchQA | Search | 1 | 94.4% | 2026-05-28 |
| SWE-bench Pro | Software Engineering | 1 | 77.8% | 2026-04-16 |
| SWE-bench Verified | Software Engineering | 1 | 93.9% | 2026-04-16 |
No matching rows.