benchgap
OpenAI · model

GPT-6.2 Astra benchmark scores

As of 2026-11-18, GPT-6.2 Astra (OpenAI) has measured scores on 34 benchmarks and estimated scores on 95 more.

BenchmarkScoreSource
AA-GPQA Diamond96.2%estimated ± 2.1 pp, low confidence
AA-HLE61.4%estimated ± 3.9 pp, medium confidence
AA-SciCode61.7%estimated ± 3.5 pp, medium confidence
AIME2698.2%estimated ± 1.4 pp, medium confidence
Artificial Analysis Intelligence Index62.1%measured
CritPt35.3%estimated ± 4.0 pp, medium confidence
GDPval-AA61.1%estimated ± 6.9 pp, low confidence
GPQA Diamond97.0%measured
HLE74.5%estimated ± 2.2 pp, medium confidence
AA-Omniscience Accuracy68.6%measured
Terminal-Bench 2.189.1%estimated ± 5.4 pp, low confidence
BrowseComp92.7%estimated ± 2.9 pp, low confidence
HLE w/ tools64.6%estimated ± 5.1 pp, low confidence
AA Coding Index79.0%estimated ± 2.3 pp, low confidence
AA-MMMU-Pro89.5%measured
BioMysteryBench (human-difficult)51.4%estimated ± 7.8 pp, low confidence
AutomationBench52.3%estimated ± 6.1 pp, low confidence
JobBench73.6%estimated ± 11.3 pp, low confidence
SWE-bench Pro85.9%estimated ± 8.4 pp, medium confidence
AA Agentic Index52.2%estimated ± 12.7 pp, low confidence
MMLU-Pro85.5%estimated ± 5.1 pp, low confidence
FailureSensorIQ69.4%estimated ± 2.1 pp, medium confidence
GPQA97.0%measured
MMLU88.1%estimated ± 4.3 pp, medium confidence
Vals GPQA Diamond98.8%estimated ± 1.2 pp, low confidence
Vals MMLU-Pro90.3%estimated ± 1.6 pp, medium confidence
MMMLU91.8%estimated ± 1.1 pp, low confidence
MMMU91.9%estimated ± 2.7 pp, medium confidence
Chess Puzzles78.0%measured
AA MMLU-Pro91.6%estimated ± 0.8 pp, medium confidence
AA-AnalystAgent62.5%measured
AA EnterpriseOps-Gym29.9%measured
AA Tau3 Banking100.0%estimated ± 2.4 pp, low confidence
ARC-AGI-198.5%measured
ARC-AGI-295.8%measured
BabaIsBench100.0%estimated ± 1.2 pp, medium confidence
CursorBench 3.282.5%estimated ± 4.1 pp, low confidence
EBR-bench92.4%measured
FrontierCode 1.1 Main51.8%estimated ± 3.9 pp, medium confidence
FrontierMath v2 (Tier 4)100.0%measured
FrontierMath v2 (Tiers 1-3)92.6%measured
FrontierSWE v260.6%estimated ± 8.3 pp, medium confidence
Furniture Assembly89.2%measured
MaxxBench76.8%estimated ± 4.3 pp, low confidence
MazeBench9.5%estimated ± 6.2 pp, low confidence
MirrorCode67.8%measured
Mystery Game Puzzles95.0%measured
NYT Connections95.0%estimated ± 6.7 pp, low confidence
OfficeQA Pro67.6%estimated ± 6.8 pp, low confidence
ParseBench-178.0%estimated ± 6.2 pp, low confidence
Terminal-Bench 4.165.0%estimated ± 5.5 pp, low confidence
Terminal-Bench 2.1 (Vals AI)84.2%estimated ± 6.9 pp, low confidence
CWE-Bench-AA65.0%measured
AA Cyber Index37.3%measured
DeepsecBench-AA46.8%measured
Harvey LAB-AA v1.115.0%measured
HealthBench Professional (length-adjusted, Artificial Analysis)64.8%estimated ± 4.3 pp, medium confidence
HealthBench Professional (raw, Artificial Analysis)71.4%estimated ± 1.1 pp, medium confidence
Bug Hunt Bench47.5%estimated ± 8.9 pp, medium confidence
Chartography (tools)88.9%estimated ± 4.1 pp, low confidence
CursorBench 4.057.6%estimated ± 3.3 pp, high confidence
Epoch Automation Reports65.7%estimated ± 3.0 pp, low confidence
FrontierCode 2.044.2%measured
FrontierPhysics20.6%estimated ± 2.3 pp, high confidence
HLE w/o tools51.5%estimated ± 4.2 pp, medium confidence
OSWorld 2.072.5%estimated ± 14.7 pp, low confidence
OSWorld 2.1 Offline98.7%estimated ± 7.9 pp, low confidence
PostTrainBench v1.146.0%estimated ± 4.8 pp, high confidence
SpireClimb41.3%measured
StealthMark68.6%estimated ± 6.1 pp, medium confidence
Terminal-Bench 4.067.7%estimated ± 6.1 pp, medium confidence
Terminal-Bench-Science 0.177.1%measured
Toolathlon-Verified79.4%estimated ± 2.8 pp, low confidence
VulcanBench Frontier v491.7%estimated ± 6.0 pp, low confidence
AA AutomationBench77.2%measured
GDP.pdf29.8%measured
AA Terminal-Bench 4.064.1%measured
ARC-AGI-370.3%measured
ARC-AGI-3 Provider Adapter99.8%measured
CWE-bench75.5%estimated ± 6.2 pp, low confidence
MKA53.2%estimated ± 5.3 pp, medium confidence
SkillsBench official leaderboard (BenchFlow)60.8%estimated ± 6.2 pp, low confidence
FACTS Multimodal52.4%estimated ± 2.7 pp, medium confidence
Hermes Bench100.0%estimated ± 1.8 pp, low confidence
Hermes Index92.4%estimated ± 2.7 pp, low confidence
ParseBench-287.0%estimated ± 6.2 pp, low confidence
CharXiv89.5%estimated ± 6.3 pp, low confidence
CharXiv w/o tools88.8%estimated ± 2.5 pp, medium confidence
ExploitBench100.0%measured
ExploitGym41.5%estimated ± 9.8 pp, low confidence
AI2D_TEST94.2%estimated ± 2.0 pp, low confidence
C-Eval93.9%estimated ± 1.2 pp, medium confidence
ERQA68.8%estimated ± 3.1 pp, medium confidence
HMMT Feb 2026100.0%estimated ± 3.1 pp, medium confidence
HMMT Nov 202596.1%estimated ± 1.5 pp, low confidence
MathVision96.9%estimated ± 3.1 pp, medium confidence
MMAnswerBench97.4%estimated ± 2.4 pp, low confidence
MMLU-Redux95.8%estimated ± 1.7 pp, medium confidence
MMMU-Pro87.8%estimated ± 1.5 pp, medium confidence
RealWorldQA91.7%estimated ± 2.9 pp, low confidence
ScreenSpot Pro82.9%estimated ± 10.8 pp, low confidence
SuperGPQA84.7%estimated ± 8.5 pp, low confidence
Video-MME (w/o subtitle)93.9%estimated ± 1.7 pp, low confidence
Video-MME (with subtitle)93.8%estimated ± 3.7 pp, low confidence
HealthBench Hard34.6%estimated ± 1.2 pp, medium confidence
MedXpertQA (MM)77.9%estimated ± 6.5 pp, low confidence
MedXpertQA (Text)83.5%estimated ± 4.5 pp, low confidence
MMLU-Pro (Arcee)89.1%estimated ± 3.4 pp, low confidence
PostTrainBench42.0%estimated ± 4.6 pp, high confidence
DAYJOB16.6%estimated ± 1.1 pp, medium confidence
DAYJOB Healthcare11.2%estimated ± 1.1 pp, low confidence
HealthBench (raw)54.7%estimated ± 1.6 pp, high confidence
HealthBench (length-adjusted)58.4%estimated ± 0.9 pp, low confidence
HealthBench Professional66.3%measured
HealthBench Professional (raw)66.0%measured
HLE-Verified54.6%estimated ± 0.7 pp, low confidence
LABBench286.1%estimated ± 1.7 pp, medium confidence
ScarfBench69.9%estimated ± 5.0 pp, low confidence
BlenderWorld (text-only)86.6%estimated ± 2.8 pp, low confidence
Cookie-Bench90.8%estimated ± 1.3 pp, low confidence
STEMBench2.7%estimated ± 0.2 pp, low confidence
SimpleQA55.0%estimated ± 14.6 pp, low confidence
Agents' Last Exam59.1%estimated ± 11.4 pp, low confidence
BabyVision w/ Python96.4%estimated ± 4.0 pp, low confidence
SEC-Bench Pro86.8%measured
BioBasicsBench98.2%estimated ± 2.2 pp, low confidence
LifeVal86.1%measured
MMMU-Pro w/ Python91.3%estimated ± 2.0 pp, medium confidence
ExploitBench Internal Port53.5%measured