benchgap
OpenAI · model

GPT-5.4 mini benchmark scores

As of 2026-11-22, GPT-5.4 mini (OpenAI) has measured scores on 41 benchmarks and estimated scores on 140 more.

BenchmarkScoreSource
AA-GPQA Diamond87.5%measured
AA-HLE28.1%measured
AA-SciCode52.1%measured
AIME2695.9%estimated ± 1.2 pp, medium confidence
Artificial Analysis Intelligence Index24.1%measured
CritPt10.0%measured
GDPval-AA25.6%measured
GPQA Diamond88.0%estimated ± 0.1 pp, medium confidence
HLE41.5%measured
IFBench71.3%estimated ± 6.6 pp, medium confidence
AA-LCR77.0%measured
LiveCodeBench v687.1%estimated ± 2.2 pp, medium confidence
AA-Omniscience Accuracy37.5%measured
SciCode39.7%estimated ± 2.9 pp, low confidence
Terminal-Bench 2.170.6%estimated ± 5.6 pp, medium confidence
BrowseComp82.3%estimated ± 4.2 pp, high confidence
HLE w/ tools56.8%estimated ± 4.1 pp, medium confidence
IFEval92.9%estimated ± 2.8 pp, high confidence
LongBench v263.2%estimated ± 5.5 pp, medium confidence
VITA-Bench38.5%estimated ± 9.9 pp, low confidence
τ²-bench results93.4%measured
AA Coding Index56.1%measured
AA-MMMU-Pro73.3%measured
APEX-Agents30.6%estimated ± 4.9 pp, low confidence
APEX-Agents-AA28.2%measured
BioMysteryBench (human-difficult)43.8%estimated ± 6.8 pp, low confidence
DeepSearchQA84.0%estimated ± 4.5 pp, high confidence
SWE-bench Verified80.5%estimated ± 2.2 pp, medium confidence
BFCL v473.5%estimated ± 2.7 pp, high confidence
CyberGym60.6%estimated ± 5.3 pp, low confidence
JobBench32.0%estimated ± 7.3 pp, low confidence
SkillsBench45.6%estimated ± 9.1 pp, low confidence
SWE-bench Pro56.8%estimated ± 2.9 pp, medium confidence
WideResearch81.4%estimated ± 2.1 pp, medium confidence
AA Agentic Index19.7%measured
MMLU-Pro85.1%estimated ± 2.5 pp, medium confidence
AA-IFBench73.3%measured
FailureSensorIQ66.0%estimated ± 1.7 pp, high confidence
GPQA88.0%measured
MMLU90.5%estimated ± 0.9 pp, medium confidence
Vals GPQA Diamond83.1%measured
Vals LiveCodeBench81.5%measured
Vals MMLU-Pro84.6%measured
MMMLU89.5%estimated ± 0.7 pp, medium confidence
MMMU82.8%estimated ± 0.8 pp, high confidence
Chess Puzzles17.0%measured
AA MMLU-Pro87.0%estimated ± 0.7 pp, high confidence
Gert Labs47.7%estimated ± 9.3 pp, medium confidence
AA Harvey LAB v1.080.8%estimated ± 1.2 pp, low confidence
AA Tau3 Banking39.6%estimated ± 2.0 pp, low confidence
AA Terminal-Bench 2.180.0%estimated ± 2.5 pp, medium confidence
ApprenticeBench3.5%estimated ± 5.0 pp, low confidence
ARC-AGI-163.7%measured
ARC-AGI-218.9%measured
CursorBench 3.149.2%estimated ± 3.4 pp, medium confidence
CursorBench 3.241.0%estimated ± 4.4 pp, medium confidence
EBR-bench4.2%estimated ± 8.1 pp, low confidence
Terminal-Bench 3.08.3%estimated ± 3.2 pp, medium confidence
FrontierCode 1.1 Main27.0%measured
FrontierMath v2 (Tier 4)9.8%measured
FrontierMath v2 (Tiers 1-3)51.2%measured
FrontierSWE v27.3%estimated ± 9.5 pp, low confidence
Furniture Assembly32.9%estimated ± 9.3 pp, medium confidence
MaxxBench47.2%estimated ± 8.8 pp, medium confidence
MazeBench0.4%estimated ± 5.5 pp, low confidence
Mystery Game Puzzles7.0%measured
NYT Connections61.8%measured
OfficeQA Pro50.3%estimated ± 6.8 pp, low confidence
OSWorld-Verified72.1%measured
ParseBench-159.0%estimated ± 5.5 pp, low confidence
Terminal-Bench 4.11.8%measured
Terminal-Bench Hard52.3%measured
Vals SWE-bench73.0%measured
Terminal-Bench 2.1 (Vals AI)54.7%measured
VulcanBench v381.6%estimated ± 5.6 pp, low confidence
CWE-Bench-AA33.1%estimated ± 10.1 pp, low confidence
AA Cyber Index13.3%estimated ± 9.2 pp, low confidence
DeepsecBench-AA1.5%estimated ± 8.9 pp, low confidence
HealthBench Professional (length-adjusted, Artificial Analysis)53.1%estimated ± 4.3 pp, low confidence
HealthBench Professional (raw, Artificial Analysis)55.1%estimated ± 0.8 pp, low confidence
AA ITBench35.2%measured
Bug Hunt Bench5.8%estimated ± 9.0 pp, low confidence
Chartography (tools)79.2%estimated ± 4.1 pp, low confidence
CursorBench 4.030.8%estimated ± 3.4 pp, medium confidence
CWE-bench v135.1%estimated ± 6.4 pp, low confidence
DeepSWE24.7%estimated ± 6.4 pp, low confidence
DRACO73.0%estimated ± 2.0 pp, medium confidence
FrontierCode 2.026.1%estimated ± 9.3 pp, low confidence
FrontierPhysics2.5%estimated ± 5.0 pp, low confidence
HLE w/o tools28.2%measured
OSWorld 2.06.2%estimated ± 3.6 pp, high confidence
OSWorld 2.1 Offline2.8%estimated ± 1.9 pp, low confidence
PostTrainBench v1.118.0%estimated ± 3.7 pp, medium confidence
SpireClimb5.8%estimated ± 7.2 pp, low confidence
StealthMark32.1%estimated ± 9.2 pp, medium confidence
Terminal-Bench 4.010.7%estimated ± 4.5 pp, medium confidence
Toolathlon-Verified47.7%estimated ± 2.3 pp, low confidence
VulcanBench Frontier v482.2%estimated ± 6.8 pp, low confidence
GDP.pdf17.0%estimated ± 4.6 pp, medium confidence
AA Terminal-Bench 4.01.4%estimated ± 6.2 pp, low confidence
ARC-AGI-30.4%estimated ± 7.2 pp, low confidence
ARC-AGI-3 Provider Adapter2.7%estimated ± 7.3 pp, low confidence
cursorBench4144.8%estimated ± 3.7 pp, medium confidence
MKA21.7%estimated ± 6.4 pp, low confidence
SkillsBench official leaderboard (BenchFlow)41.4%measured
BinaryAudit25.6%estimated ± 12.0 pp, low confidence
CompileBench55.6%estimated ± 10.0 pp, low confidence
OTelBench22.5%estimated ± 6.6 pp, medium confidence
Vibe Code Bench48.0%measured
FACTS Multimodal35.5%estimated ± 4.0 pp, high confidence
Hermes Bench56.3%estimated ± 2.9 pp, high confidence
Hermes Index23.4%estimated ± 8.7 pp, medium confidence
ParseBench-267.1%estimated ± 5.0 pp, medium confidence
CharXiv78.2%estimated ± 5.6 pp, medium confidence
CharXiv w/o tools80.7%estimated ± 1.2 pp, high confidence
ExploitBench14.6%estimated ± 7.2 pp, low confidence
ExploitGym1.0%estimated ± 10.6 pp, low confidence
AI2D_TEST89.3%estimated ± 2.0 pp, medium confidence
C-Eval92.8%estimated ± 1.0 pp, medium confidence
Claw-Eval64.5%estimated ± 6.2 pp, low confidence
CountBench93.8%estimated ± 2.9 pp, medium confidence
DeepPlanning38.8%estimated ± 13.1 pp, low confidence
DynaMath84.5%estimated ± 3.0 pp, medium confidence
ERQA55.9%estimated ± 3.1 pp, high confidence
HMMT Feb 202687.2%estimated ± 1.5 pp, medium confidence
HMMT Nov 202594.5%estimated ± 1.5 pp, medium confidence
MathVision85.8%estimated ± 3.0 pp, high confidence
MCP Atlas57.7%measured
MCP-Tasks76.2%estimated ± 0.6 pp, low confidence
MLVU (M-Avg)85.3%estimated ± 2.2 pp, high confidence
MMAnswerBench86.9%estimated ± 2.4 pp, medium confidence
MMLU-Redux95.7%estimated ± 0.8 pp, medium confidence
MMMU-Pro76.6%measured
NL2Repo41.5%estimated ± 4.7 pp, high confidence
PostTrainBench v119.6%estimated ± 3.1 pp, medium confidence
QwenClawBench55.2%estimated ± 2.6 pp, high confidence
RealWorldQA83.4%estimated ± 2.8 pp, high confidence
ScreenSpot Pro74.5%estimated ± 9.6 pp, low confidence
SimpleVQA64.6%estimated ± 7.5 pp, medium confidence
SuperGPQA81.3%estimated ± 7.2 pp, medium confidence
Terminal-Bench 2.060.0%measured
Toolathlon42.9%measured
V*100.0%estimated ± 7.5 pp, medium confidence
Video-MME (w/o subtitle)83.7%estimated ± 1.3 pp, high confidence
Video-MME (with subtitle)86.2%estimated ± 3.0 pp, high confidence
VideoMMMU84.5%estimated ± 2.4 pp, high confidence
We-Math81.6%estimated ± 7.4 pp, low confidence
AA LiveCodeBench88.0%estimated ± 2.6 pp, high confidence
HealthBench Hard0.0%estimated ± 5.1 pp, low confidence
MedXpertQA (MM)67.0%estimated ± 2.4 pp, high confidence
MedXpertQA (Text)44.3%estimated ± 8.9 pp, low confidence
React Native Evals81.1%estimated ± 3.0 pp, medium confidence
ResearchClawBench17.1%estimated ± 1.8 pp, high confidence
SWE-Rebench62.0%estimated ± 6.5 pp, low confidence
PostTrainBench15.4%estimated ± 3.7 pp, medium confidence
DAYJOB1.3%estimated ± 3.2 pp, high confidence
DAYJOB Healthcare0.7%estimated ± 4.2 pp, high confidence
BioMysteryBench (human-solvable)83.3%estimated ± 0.6 pp, low confidence
HealthBench (raw)36.9%estimated ± 6.0 pp, low confidence
HealthBench (length-adjusted)49.1%estimated ± 3.4 pp, low confidence
HealthBench Professional38.2%estimated ± 4.2 pp, low confidence
HealthBench Professional (raw)48.2%estimated ± 4.5 pp, medium confidence
HLE-Verified15.3%estimated ± 2.0 pp, low confidence
LABBench280.1%estimated ± 1.8 pp, low confidence
ScarfBench25.2%estimated ± 9.1 pp, low confidence
BlenderWorld (text-only)70.0%estimated ± 2.8 pp, low confidence
Cookie-Bench76.9%estimated ± 1.3 pp, low confidence
STEMBench2.5%estimated ± 0.2 pp, medium confidence
SimpleQA42.1%estimated ± 4.3 pp, low confidence
BabyVision w/ Python88.8%estimated ± 3.8 pp, low confidence
OpenHarmony Bench v1.058.8%estimated ± 2.4 pp, high confidence
SEC-Bench Pro33.7%estimated ± 9.3 pp, medium confidence
BabyVision57.5%estimated ± 9.1 pp, low confidence
MMVU76.3%estimated ± 2.7 pp, medium confidence
PerceptionBench58.9%estimated ± 4.9 pp, low confidence
MRCRv285.2%estimated ± 11.3 pp, low confidence
MMLongBench-Doc61.3%estimated ± 2.6 pp, low confidence
OCRBench V270.8%estimated ± 3.3 pp, high confidence
BioBasicsBench33.9%estimated ± 2.8 pp, low confidence
MMMU-Pro w/ Python78.0%measured
ExploitBench Internal Port0.0%estimated ± 1.4 pp, low confidence