benchgap
xAI · model

Grok 4.20 benchmark scores

As of 2026-11-22, Grok 4.20 (xAI) has measured scores on 26 benchmarks and estimated scores on 128 more.

BenchmarkScoreSource
AA-GPQA Diamond90.0%estimated ± 2.7 pp, high confidence
AA-HLE31.5%estimated ± 3.9 pp, high confidence
AA-SciCode46.0%estimated ± 3.6 pp, high confidence
Artificial Analysis Intelligence Index21.1%estimated ± 5.6 pp, medium confidence
CritPt8.3%estimated ± 4.6 pp, high confidence
GDPval-AA16.6%estimated ± 10.5 pp, low confidence
GPQA Diamond88.5%measured
HLE54.2%estimated ± 2.2 pp, medium confidence
LiveCodeBench v688.4%estimated ± 2.2 pp, medium confidence
AA-Omniscience Accuracy37.3%estimated ± 6.5 pp, medium confidence
SciCode45.3%estimated ± 2.9 pp, medium confidence
Terminal-Bench 2.146.8%estimated ± 6.7 pp, low confidence
BrowseComp76.3%estimated ± 6.8 pp, low confidence
HLE w/ tools42.9%estimated ± 5.7 pp, medium confidence
AA Coding Index54.0%estimated ± 5.3 pp, medium confidence
AA-MMMU-Pro74.4%estimated ± 1.5 pp, high confidence
APEX-Agents-AA10.3%estimated ± 9.3 pp, medium confidence
DeepSearchQA62.8%measured
SWE-bench Verified76.7%measured
CyberGym44.0%estimated ± 11.0 pp, low confidence
JobBench16.0%estimated ± 14.2 pp, low confidence
SWE-bench Pro51.8%measured
WideResearch64.3%estimated ± 3.3 pp, high confidence
AA Agentic Index13.2%estimated ± 9.4 pp, medium confidence
MMLU-Pro85.3%estimated ± 2.5 pp, medium confidence
FailureSensorIQ67.7%estimated ± 1.7 pp, high confidence
GPQA88.5%estimated ± 0.0 pp, medium confidence
MMLU93.6%estimated ± 1.2 pp, medium confidence
Vals GPQA Diamond88.6%measured
Vals LiveCodeBench84.3%measured
Vals MMLU-Pro86.3%measured
MMMLU89.3%estimated ± 1.1 pp, medium confidence
MMMU82.8%estimated ± 1.7 pp, high confidence
Chess Puzzles29.0%measured
AA MMLU-Pro87.3%estimated ± 1.1 pp, high confidence
Gert Labs22.3%measured
AA-AnalystAgent44.8%estimated ± 5.9 pp, low confidence
AA EnterpriseOps-Gym40.9%estimated ± 2.3 pp, low confidence
AA Terminal-Bench 2.179.4%estimated ± 2.5 pp, medium confidence
ApprenticeBench3.8%estimated ± 10.1 pp, low confidence
ARC-AGI-186.7%estimated ± 1.7 pp, high confidence
ARC-AGI-253.3%measured
CursorBench 3.149.0%estimated ± 5.6 pp, low confidence
CursorBench 3.244.1%estimated ± 3.6 pp, medium confidence
EBR-bench15.5%estimated ± 12.7 pp, low confidence
FrontierCode 1.1 Main21.7%estimated ± 4.2 pp, low confidence
FrontierSWE v21.4%estimated ± 8.3 pp, low confidence
Furniture Assembly29.9%estimated ± 14.1 pp, low confidence
MaxxBench46.0%estimated ± 10.3 pp, low confidence
MazeBench0.0%measured
Mystery Game Puzzles20.6%estimated ± 10.3 pp, low confidence
NYT Connections85.4%measured
OSWorld-Verified63.2%estimated ± 7.1 pp, low confidence
ParseBench-157.3%estimated ± 8.4 pp, medium confidence
Terminal-Bench 4.16.2%estimated ± 14.9 pp, low confidence
Terminal-Bench Hard44.1%estimated ± 3.9 pp, medium confidence
Vals SWE-bench72.2%measured
Terminal-Bench 2.1 (Vals AI)44.2%measured
VulcanBench v383.2%estimated ± 5.6 pp, low confidence
CWE-Bench-AA33.4%estimated ± 8.1 pp, low confidence
AA Cyber Index5.8%estimated ± 12.4 pp, low confidence
DeepsecBench-AA6.5%estimated ± 8.2 pp, low confidence
HealthBench Professional (length-adjusted, Artificial Analysis)57.5%estimated ± 4.4 pp, low confidence
HealthBench Professional (raw, Artificial Analysis)59.3%estimated ± 2.6 pp, low confidence
AA ITBench35.7%estimated ± 4.6 pp, low confidence
Bug Hunt Bench27.6%estimated ± 12.1 pp, low confidence
CursorBench 4.029.4%estimated ± 3.3 pp, medium confidence
CWE-bench v157.4%estimated ± 5.1 pp, medium confidence
DeepSWE56.3%estimated ± 7.2 pp, low confidence
DRACO86.9%estimated ± 0.7 pp, low confidence
FrontierCode 2.019.2%estimated ± 6.7 pp, low confidence
FrontierPhysics0.0%estimated ± 6.1 pp, low confidence
HLE w/o tools31.6%measured
OSWorld 2.12.2%estimated ± 3.6 pp, medium confidence
OSWorld 2.1 Offline33.3%estimated ± 13.8 pp, low confidence
PostTrainBench v1.117.3%estimated ± 3.7 pp, medium confidence
SpireClimb5.8%estimated ± 7.2 pp, medium confidence
StealthMark42.1%estimated ± 9.7 pp, low confidence
Toolathlon-Verified73.4%estimated ± 2.6 pp, high confidence
VulcanBench Frontier v474.7%estimated ± 8.3 pp, low confidence
AA AutomationBench39.0%estimated ± 9.9 pp, low confidence
GDP.pdf10.5%estimated ± 4.2 pp, medium confidence
AA Terminal-Bench 4.00.0%estimated ± 13.7 pp, low confidence
ARC-AGI-30.1%measured
ARC-AGI-3 Provider Adapter0.5%estimated ± 4.4 pp, medium confidence
CWE-bench27.4%estimated ± 6.2 pp, low confidence
MKA32.8%estimated ± 4.6 pp, medium confidence
SkillsBench official leaderboard (BenchFlow)54.1%estimated ± 6.1 pp, medium confidence
BinaryAudit35.8%estimated ± 12.1 pp, low confidence
OTelBench22.5%estimated ± 6.6 pp, medium confidence
Vibe Code Bench4.1%measured
FACTS Multimodal48.0%estimated ± 4.0 pp, high confidence
Hermes Index37.2%estimated ± 3.1 pp, high confidence
ParseBench-265.1%estimated ± 5.0 pp, medium confidence
CharXiv60.9%measured
CharXiv w/o tools79.5%estimated ± 1.2 pp, high confidence
ExploitBench31.1%estimated ± 13.4 pp, low confidence
ExploitGym11.9%estimated ± 6.3 pp, medium confidence
AI2D_TEST87.8%estimated ± 0.9 pp, low confidence
CountBench93.3%estimated ± 2.2 pp, medium confidence
DynaMath72.3%estimated ± 0.8 pp, low confidence
ERQA54.1%measured
MathVision84.2%estimated ± 3.0 pp, high confidence
MLVU (M-Avg)79.4%estimated ± 1.5 pp, medium confidence
MMLU-Redux92.4%estimated ± 2.8 pp, high confidence
MMMU-Pro75.2%measured
NL2Repo41.1%estimated ± 4.7 pp, high confidence
OmniDocBench 1.587.5%estimated ± 2.0 pp, medium confidence
PostTrainBench v119.0%estimated ± 3.1 pp, low confidence
RealWorldQA74.1%estimated ± 1.5 pp, medium confidence
ScreenSpot Pro31.1%estimated ± 6.8 pp, low confidence
SimpleVQA57.4%measured
SuperGPQA73.9%estimated ± 13.9 pp, low confidence
Terminal-Bench 2.047.1%measured
Toolathlon37.8%estimated ± 8.0 pp, medium confidence
V*81.3%estimated ± 11.2 pp, low confidence
Video-MME (w/o subtitle)83.0%estimated ± 1.3 pp, high confidence
Video-MME (with subtitle)76.5%estimated ± 2.9 pp, medium confidence
VideoMMMU83.6%estimated ± 1.5 pp, low confidence
We-Math75.2%estimated ± 2.1 pp, medium confidence
AA LiveCodeBench87.6%estimated ± 2.6 pp, high confidence
HealthBench Hard20.3%measured
LiveCodeBench Pro74.2%measured
MedXpertQA (MM)65.8%measured
MedXpertQA (Text)50.2%measured
MMLU-Pro (Arcee)86.4%estimated ± 3.4 pp, medium confidence
React Native Evals78.2%estimated ± 2.7 pp, medium confidence
ResearchClawBench13.6%estimated ± 2.9 pp, low confidence
SWE-Rebench59.0%estimated ± 3.6 pp, medium confidence
PostTrainBench14.7%estimated ± 3.7 pp, medium confidence
DAYJOB0.5%estimated ± 6.3 pp, medium confidence
DAYJOB Healthcare0.2%estimated ± 6.1 pp, low confidence
Finance Agent v255.1%estimated ± 2.4 pp, low confidence
FrontierCode 1.1 Extended53.0%estimated ± 3.0 pp, low confidence
HealthBench (length-adjusted)53.8%estimated ± 0.9 pp, low confidence
HealthBench Professional56.8%estimated ± 3.6 pp, medium confidence
HLE-Verified26.0%estimated ± 4.6 pp, low confidence
LABBench281.1%estimated ± 1.8 pp, low confidence
ScarfBench43.2%estimated ± 12.3 pp, low confidence
BlenderWorld (text-only)70.6%estimated ± 4.0 pp, medium confidence
Cookie-Bench79.7%estimated ± 2.5 pp, medium confidence
LiveCodeBench77.3%estimated ± 14.6 pp, low confidence
Agents' Last Exam29.6%estimated ± 9.7 pp, low confidence
BabyVision w/ Python88.8%estimated ± 3.8 pp, low confidence
OpenHarmony Bench v1.051.8%estimated ± 3.2 pp, high confidence
SEC-Bench Pro41.3%estimated ± 12.3 pp, low confidence
BabyVision54.2%estimated ± 9.1 pp, low confidence
MMVU75.5%estimated ± 2.7 pp, medium confidence
PerceptionBench58.1%estimated ± 4.9 pp, low confidence
OCRBench V270.4%estimated ± 3.3 pp, high confidence
BioBasicsBench36.8%estimated ± 3.5 pp, low confidence
LifeVal75.2%estimated ± 3.8 pp, medium confidence
MMMU-Pro w/ Python78.8%estimated ± 2.6 pp, high confidence
ExploitBench Internal Port0.1%estimated ± 1.4 pp, low confidence