benchgap
Anthropic · model

Claude Mythos 5.1 benchmark scores

As of 2026-11-22, Claude Mythos 5.1 (Anthropic) has measured scores on 2 benchmarks and estimated scores on 19 more.

BenchmarkScoreSource
GDPval-AA55.3%estimated ± 7.6 pp, low confidence
Terminal-Bench 2.188.2%estimated ± 6.5 pp, low confidence
BrowseComp88.2%estimated ± 3.0 pp, medium confidence
CyberGym86.0%estimated ± 10.0 pp, medium confidence
AA Agentic Index45.4%estimated ± 8.0 pp, low confidence
Gert Labs69.7%estimated ± 7.7 pp, medium confidence
AA-AnalystAgent44.2%estimated ± 5.5 pp, low confidence
MazeBench5.3%estimated ± 3.7 pp, high confidence
Terminal-Bench 4.160.5%estimated ± 4.6 pp, high confidence
CWE-bench v157.8%estimated ± 4.5 pp, medium confidence
Terminal-Bench 4.060.9%measured
Terminal-Bench-Science 0.157.9%estimated ± 14.2 pp, low confidence
AA AutomationBench68.2%estimated ± 4.8 pp, medium confidence
GDP.pdf24.0%estimated ± 7.4 pp, low confidence
AA Terminal-Bench 4.057.1%estimated ± 6.3 pp, medium confidence
CWE-bench55.9%estimated ± 8.2 pp, low confidence
MKA44.9%estimated ± 3.1 pp, high confidence
Hermes Bench73.6%estimated ± 1.5 pp, medium confidence
Hermes Index48.7%estimated ± 5.5 pp, low confidence
ExploitGym23.4%measured
Agents' Last Exam51.7%estimated ± 9.6 pp, medium confidence