benchgap
Anthropic · model

Claude Fable 5.2 benchmark scores

As of 2026-11-18, Claude Fable 5.2 (Anthropic) has measured scores on 47 benchmarks and estimated scores on 98 more.

BenchmarkScoreSource
AA-GPQA Diamond95.3%estimated ± 2.7 pp, medium confidence
AA-HLE74.3%measured
AA-SciCode67.7%measured
AIME2698.2%estimated ± 1.4 pp, medium confidence
Artificial Analysis Intelligence Index64.6%measured
CritPt35.4%measured
GDPval-AA70.6%estimated ± 1.7 pp, low confidence
GPQA Diamond96.6%estimated ± 1.8 pp, medium confidence
HLE74.2%measured
AA-LCR83.3%measured
LiveCodeBench v6100.0%estimated ± 2.7 pp, medium confidence
AA-Omniscience Accuracy79.1%measured
SciCode67.0%estimated ± 3.1 pp, medium confidence
Terminal-Bench 2.189.1%estimated ± 5.4 pp, low confidence
BrowseComp91.0%estimated ± 2.7 pp, medium confidence
HLE w/ tools64.2%estimated ± 6.8 pp, low confidence
LongBench v268.6%estimated ± 5.5 pp, medium confidence
AA Coding Index86.4%estimated ± 1.2 pp, medium confidence
AA-MMMU-Pro100.0%estimated ± 2.3 pp, medium confidence
APEX-Agents51.6%estimated ± 7.4 pp, low confidence
BioMysteryBench (human-difficult)50.2%estimated ± 6.8 pp, low confidence
SWE-bench Verified100.0%estimated ± 3.6 pp, medium confidence
AutomationBench51.4%measured
CyberGym87.4%estimated ± 4.6 pp, medium confidence
JobBench69.2%estimated ± 11.3 pp, low confidence
SWE-bench Pro93.0%measured
MMLU-Pro86.2%estimated ± 2.5 pp, low confidence
FailureSensorIQ68.6%estimated ± 1.7 pp, high confidence
GPQA96.5%estimated ± 0.9 pp, high confidence
Vals GPQA Diamond96.5%estimated ± 2.9 pp, medium confidence
Vals LiveCodeBench94.6%estimated ± 3.0 pp, medium confidence
Vals MMLU-Pro92.8%estimated ± 1.8 pp, medium confidence
MMMLU95.5%estimated ± 2.6 pp, low confidence
Chess Puzzles79.0%measured
AA MMLU-Pro93.2%estimated ± 0.8 pp, medium confidence
AA-AnalystAgent58.1%estimated ± 2.7 pp, low confidence
AA Tau3 Banking50.0%estimated ± 3.3 pp, low confidence
ApprenticeBench60.8%estimated ± 12.5 pp, low confidence
ARC-AGI-197.8%estimated ± 3.4 pp, high confidence
ARC-AGI-292.7%estimated ± 9.4 pp, medium confidence
BabaIsBench100.0%estimated ± 1.2 pp, medium confidence
CursorBench 3.187.6%estimated ± 3.4 pp, low confidence
CursorBench 3.277.2%estimated ± 2.0 pp, medium confidence
EBR-bench90.5%measured
Terminal-Bench 3.050.4%estimated ± 3.2 pp, medium confidence
FrontierCode 1.1 Main52.1%estimated ± 3.5 pp, low confidence
FrontierMath v2 (Tier 4)100.0%measured
FrontierMath v2 (Tiers 1-3)95.1%measured
FrontierSWE v269.9%measured
Furniture Assembly90.0%measured
MaxxBench76.6%estimated ± 4.3 pp, low confidence
MazeBench21.0%measured
Mystery Game Puzzles90.0%measured
NYT Connections97.3%measured
OfficeQA Pro72.9%measured
ParseBench-186.6%estimated ± 4.6 pp, low confidence
Terminal-Bench 4.165.8%measured
Terminal-Bench Hard70.7%estimated ± 2.8 pp, low confidence
Vals SWE-bench100.0%estimated ± 3.7 pp, medium confidence
Terminal-Bench 2.1 (Vals AI)82.6%estimated ± 6.9 pp, low confidence
VulcanBench v387.6%estimated ± 3.9 pp, low confidence
CWE-Bench-AA62.9%estimated ± 4.1 pp, medium confidence
AA Cyber Index30.5%estimated ± 5.3 pp, low confidence
DeepsecBench-AA33.4%estimated ± 6.1 pp, low confidence
Harvey LAB-AA v1.17.2%measured
HealthBench Professional (length-adjusted, Artificial Analysis)67.7%measured
HealthBench Professional (raw, Artificial Analysis)71.1%measured
BenchCAD Vision2Code (no tools)0.8%estimated ± 0.1 pp, low confidence
BenchCAD Vision2Code (tools)1.0%estimated ± 0.0 pp, low confidence
Bug Hunt Bench58.0%measured
Chartography (no tools)73.4%measured
Chartography (tools)91.6%measured
CursorBench 4.060.0%measured
CWE-bench v168.9%estimated ± 5.3 pp, medium confidence
DeepSWE74.3%measured
DRACO87.7%measured
Epoch Automation Reports65.7%estimated ± 3.0 pp, low confidence
FrontierCode 2.050.2%measured
FrontierPhysics50.5%estimated ± 3.2 pp, medium confidence
HLE w/o tools73.2%measured
OfficeQA83.4%measured
OSWorld 2.178.6%estimated ± 6.1 pp, medium confidence
OSWorld 2.1 Offline89.3%measured
PostTrainBench v1.149.3%estimated ± 1.7 pp, low confidence
ProgramBench85.6%measured
SpireClimb15.3%measured
StealthMark74.7%estimated ± 6.5 pp, low confidence
SWE Multilingual96.6%measured
SWE-bench Multimodal74.1%measured
Terminal-Bench 4.067.6%measured
Terminal-Bench-Science 0.174.6%measured
Toolathlon-Verified78.4%estimated ± 1.2 pp, medium confidence
VulcanBench Frontier v493.9%measured
AA AutomationBench76.4%measured
GDP.pdf24.2%measured
AA Terminal-Bench 4.059.6%measured
ARC-AGI-365.7%measured
ARC-AGI-3 Provider Adapter94.9%measured
CWE-bench75.0%measured
MKA56.6%measured
SkillsBench official leaderboard (BenchFlow)60.3%estimated ± 6.1 pp, low confidence
BinaryAudit35.8%estimated ± 12.1 pp, low confidence
OTelBench22.5%estimated ± 6.6 pp, low confidence
FACTS Multimodal54.2%estimated ± 4.0 pp, high confidence
Hermes Bench71.9%estimated ± 1.8 pp, medium confidence
Hermes Index38.2%estimated ± 2.7 pp, medium confidence
CharXiv91.7%estimated ± 5.1 pp, low confidence
ExploitBench100.0%estimated ± 6.4 pp, low confidence
ExploitGym43.5%estimated ± 6.4 pp, medium confidence
HMMT Feb 2026100.0%estimated ± 3.1 pp, medium confidence
HMMT Nov 202596.2%estimated ± 1.5 pp, low confidence
INCLUDE91.1%estimated ± 2.0 pp, low confidence
MCP Atlas85.0%estimated ± 7.6 pp, low confidence
MMAnswerBench98.1%estimated ± 2.4 pp, low confidence
NL2Repo62.7%estimated ± 6.0 pp, medium confidence
SuperGPQA100.0%estimated ± 7.2 pp, low confidence
HealthBench Hard34.4%estimated ± 8.6 pp, low confidence
LiveCodeBench Pro100.0%estimated ± 6.0 pp, low confidence
React Native Evals84.2%estimated ± 4.7 pp, low confidence
PostTrainBench42.2%estimated ± 2.4 pp, medium confidence
DAYJOB19.1%estimated ± 2.7 pp, medium confidence
DAYJOB Healthcare18.2%estimated ± 4.0 pp, medium confidence
Finance Agent v262.3%estimated ± 3.8 pp, low confidence
BioMysteryBench (human-solvable)89.5%estimated ± 0.6 pp, low confidence
FrontierCode 1.1 Extended64.0%estimated ± 3.0 pp, low confidence
HealthBench (raw)74.8%estimated ± 6.0 pp, low confidence
HealthBench (length-adjusted)58.4%estimated ± 0.9 pp, low confidence
HealthBench Professional63.7%estimated ± 2.1 pp, medium confidence
HealthBench Professional (raw)66.4%estimated ± 2.3 pp, medium confidence
HLE-Verified90.9%estimated ± 7.1 pp, low confidence
LABBench290.7%estimated ± 2.5 pp, low confidence
ScarfBench68.9%estimated ± 3.8 pp, medium confidence
BlenderWorld (text-only)86.5%estimated ± 2.8 pp, low confidence
Cookie-Bench90.8%estimated ± 1.3 pp, low confidence
STEMBench2.8%estimated ± 0.2 pp, low confidence
Agents' Last Exam52.3%estimated ± 10.7 pp, low confidence
OpenHarmony Bench v1.058.9%estimated ± 2.4 pp, medium confidence
OpenHarmony Bench v1.263.8%estimated ± 11.0 pp, low confidence
SEC-Bench Pro76.9%estimated ± 6.9 pp, low confidence
MRCRv285.9%estimated ± 11.3 pp, low confidence
MMLongBench-Doc61.5%estimated ± 2.6 pp, low confidence
SWE-Marathon68.3%estimated ± 2.1 pp, low confidence
BioBasicsBench69.0%estimated ± 3.5 pp, low confidence
ExploitBench Internal Port57.0%estimated ± 1.4 pp, low confidence
SWE-Atlas Codebase QnA61.5%estimated ± 4.8 pp, high confidence