benchgap
Anthropic · model

Claude Mythos 5 benchmark scores

As of 2026-11-21, Claude Mythos 5 (Anthropic) has measured scores on 14 benchmarks and estimated scores on 126 more.

BenchmarkScoreSource
AA-GPQA Diamond93.6%estimated ± 2.7 pp, high confidence
AA-HLE57.9%estimated ± 4.6 pp, high confidence
AA-SciCode60.8%estimated ± 3.6 pp, high confidence
Artificial Analysis Intelligence Index51.0%estimated ± 5.6 pp, medium confidence
CritPt24.5%estimated ± 4.0 pp, high confidence
GDPval-AA50.8%estimated ± 5.0 pp, high confidence
GPQA Diamond94.1%estimated ± 1.2 pp, high confidence
HLE64.5%measured
LiveCodeBench v696.9%estimated ± 2.7 pp, medium confidence
AA-Omniscience Accuracy60.9%estimated ± 7.7 pp, medium confidence
SciCode54.3%estimated ± 3.6 pp, medium confidence
Terminal-Bench 2.188.0%measured
BrowseComp88.0%measured
HLE w/ tools61.5%estimated ± 4.9 pp, medium confidence
VITA-Bench46.8%estimated ± 6.7 pp, low confidence
τ²-bench results93.5%estimated ± 11.8 pp, low confidence
AA Coding Index78.2%estimated ± 5.3 pp, medium confidence
AA-MMMU-Pro84.7%estimated ± 2.9 pp, high confidence
APEX-Agents-AA37.7%estimated ± 4.4 pp, high confidence
DeepSearchQA90.7%estimated ± 6.8 pp, low confidence
SWE-bench Verified95.5%measured
AutomationBench45.1%estimated ± 10.4 pp, low confidence
CyberGym83.8%measured
JobBench52.4%estimated ± 8.6 pp, low confidence
SkillsBench60.6%estimated ± 4.1 pp, medium confidence
SWE-bench Pro80.3%measured
WideResearch83.0%estimated ± 1.6 pp, medium confidence
AA Agentic Index48.5%estimated ± 5.8 pp, medium confidence
MMLU-Pro86.0%estimated ± 2.5 pp, low confidence
GPQA94.1%measured
MMLU88.1%estimated ± 4.2 pp, medium confidence
Vals GPQA Diamond94.4%estimated ± 2.9 pp, high confidence
Vals LiveCodeBench88.3%estimated ± 4.7 pp, medium confidence
Vals MMLU-Pro91.3%estimated ± 1.8 pp, high confidence
MMMLU91.5%estimated ± 1.7 pp, medium confidence
MMMU93.1%estimated ± 3.7 pp, low confidence
Chess Puzzles54.4%estimated ± 8.0 pp, medium confidence
AA MMLU-Pro87.3%estimated ± 1.2 pp, low confidence
Gert Labs69.4%estimated ± 5.5 pp, low confidence
AA Harvey LAB v1.092.4%estimated ± 9.8 pp, low confidence
AA Tau3 Banking41.5%estimated ± 3.3 pp, medium confidence
AA Terminal-Bench 2.186.2%estimated ± 2.6 pp, high confidence
ApprenticeBench24.1%estimated ± 4.0 pp, high confidence
ARC-AGI-196.9%estimated ± 3.9 pp, high confidence
ARC-AGI-279.2%estimated ± 10.9 pp, low confidence
CursorBench 3.166.9%estimated ± 5.6 pp, low confidence
CursorBench 3.271.4%estimated ± 3.6 pp, high confidence
Terminal-Bench 3.023.0%estimated ± 9.1 pp, low confidence
FrontierCode 1.1 Main54.9%estimated ± 4.2 pp, medium confidence
FrontierSWE v250.7%estimated ± 8.3 pp, medium confidence
Furniture Assembly66.3%estimated ± 6.5 pp, medium confidence
MazeBench5.3%estimated ± 6.4 pp, low confidence
MirrorCode66.3%estimated ± 14.9 pp, low confidence
Mystery Game Puzzles54.8%estimated ± 7.5 pp, medium confidence
NYT Connections94.2%estimated ± 10.3 pp, low confidence
OfficeQA Pro67.0%estimated ± 2.5 pp, medium confidence
OSWorld-Verified85.0%measured
ParseBench-171.2%estimated ± 1.0 pp, low confidence
Terminal-Bench 4.124.9%estimated ± 10.9 pp, low confidence
Vals SWE-bench97.0%estimated ± 6.6 pp, medium confidence
Terminal-Bench 2.1 (Vals AI)73.7%estimated ± 7.3 pp, medium confidence
CWE-Bench-AA54.3%estimated ± 6.4 pp, medium confidence
AA Cyber Index41.6%estimated ± 13.5 pp, low confidence
DeepsecBench-AA21.9%estimated ± 9.0 pp, low confidence
HealthBench Professional (length-adjusted, Artificial Analysis)60.2%estimated ± 4.4 pp, medium confidence
HealthBench Professional (raw, Artificial Analysis)69.4%estimated ± 2.6 pp, medium confidence
AA ITBench49.2%estimated ± 4.1 pp, high confidence
Bug Hunt Bench39.7%estimated ± 12.1 pp, low confidence
CursorBench 4.050.9%estimated ± 3.3 pp, high confidence
DeepSWE64.6%estimated ± 11.2 pp, low confidence
DRACO84.1%estimated ± 2.9 pp, medium confidence
FrontierCode 2.034.4%estimated ± 6.7 pp, medium confidence
FrontierPhysics7.0%estimated ± 2.7 pp, high confidence
HLE w/o tools59.0%measured
OfficeQA78.4%estimated ± 2.7 pp, medium confidence
OSWorld 2.019.9%estimated ± 3.6 pp, high confidence
PostTrainBench v1.140.2%estimated ± 4.4 pp, high confidence
SpireClimb3.5%estimated ± 8.2 pp, low confidence
StealthMark56.5%estimated ± 6.1 pp, medium confidence
SWE Multilingual92.2%measured
SWE-bench Multimodal54.9%measured
Toolathlon-Verified73.5%estimated ± 2.2 pp, low confidence
VulcanBench Frontier v489.2%estimated ± 8.3 pp, low confidence
AA Terminal-Bench 4.038.1%estimated ± 11.0 pp, low confidence
ARC-AGI-326.5%estimated ± 7.2 pp, medium confidence
Vibe Code Bench66.3%estimated ± 8.4 pp, low confidence
FACTS Multimodal47.2%estimated ± 2.7 pp, medium confidence
CharXiv93.5%measured
CharXiv w/o tools88.9%measured
ExploitBench78.0%measured
ExploitGym22.5%estimated ± 9.2 pp, low confidence
AI2D_TEST93.5%estimated ± 0.9 pp, low confidence
C-Eval93.5%estimated ± 1.2 pp, medium confidence
Claw-Eval71.9%estimated ± 6.2 pp, low confidence
CountBench99.4%estimated ± 3.7 pp, low confidence
DynaMath88.5%estimated ± 0.8 pp, low confidence
ERQA78.9%estimated ± 2.9 pp, low confidence
INCLUDE89.8%estimated ± 2.0 pp, low confidence
MathVision93.1%estimated ± 3.3 pp, high confidence
MCP Atlas79.4%estimated ± 10.3 pp, low confidence
MLVU (M-Avg)90.1%estimated ± 1.5 pp, high confidence
MMLU-Redux95.4%estimated ± 1.7 pp, medium confidence
MMMU-Pro82.3%estimated ± 3.1 pp, high confidence
NL2Repo71.7%estimated ± 6.1 pp, low confidence
PostTrainBench v146.4%estimated ± 4.6 pp, medium confidence
RealWorldQA88.5%estimated ± 1.5 pp, high confidence
ScreenSpot Pro88.5%estimated ± 6.8 pp, medium confidence
SuperGPQA100.0%estimated ± 7.2 pp, low confidence
Toolathlon58.3%estimated ± 4.2 pp, high confidence
Video-MME (with subtitle)92.1%estimated ± 2.9 pp, high confidence
VideoMMMU88.2%estimated ± 1.5 pp, medium confidence
We-Math84.6%estimated ± 7.9 pp, low confidence
HealthBench Hard36.4%estimated ± 5.1 pp, low confidence
LiveCodeBench Pro97.4%estimated ± 6.0 pp, low confidence
MedXpertQA (MM)80.4%estimated ± 6.4 pp, low confidence
MedXpertQA (Text)74.6%estimated ± 8.9 pp, low confidence
React Native Evals86.6%estimated ± 2.7 pp, low confidence
ResearchClawBench20.3%estimated ± 2.1 pp, low confidence
SWE-Rebench61.8%estimated ± 3.6 pp, low confidence
PostTrainBench37.1%estimated ± 3.7 pp, high confidence
DAYJOB5.3%estimated ± 1.1 pp, medium confidence
DAYJOB Healthcare1.5%estimated ± 1.5 pp, medium confidence
FrontierCode 1.1 Extended61.5%estimated ± 3.0 pp, medium confidence
HealthBench Professional64.5%estimated ± 4.2 pp, medium confidence
SimpleQA38.6%estimated ± 13.4 pp, low confidence
LiveCodeBench87.9%estimated ± 14.6 pp, low confidence
Agents' Last Exam28.1%estimated ± 3.3 pp, medium confidence
BabyVision w/ Python92.1%estimated ± 3.7 pp, low confidence
OpenHarmony Bench v1.060.9%estimated ± 4.9 pp, low confidence
SEC-Bench Pro78.4%estimated ± 3.9 pp, high confidence
BabyVision85.1%estimated ± 7.4 pp, low confidence
MMVU82.1%estimated ± 3.6 pp, medium confidence
PerceptionBench66.2%estimated ± 3.8 pp, medium confidence
ZeroBench29.8%estimated ± 7.1 pp, low confidence
LVBench86.6%estimated ± 8.2 pp, low confidence
BioBasicsBench41.0%estimated ± 2.2 pp, medium confidence
LifeVal78.3%estimated ± 1.6 pp, medium confidence
ExploitBench Internal Port11.9%estimated ± 12.3 pp, low confidence
MathVision w/ Python97.7%estimated ± 1.0 pp, low confidence
PinchBench90.6%estimated ± 3.6 pp, low confidence