benchgap
Google · model

Gemini 3 Flash benchmark scores

As of 2026-11-21, Gemini 3 Flash (Google) has measured scores on 31 benchmarks and estimated scores on 119 more.

BenchmarkScoreSource
AA-GPQA Diamond81.2%measured
AA-HLE15.0%measured
AA-SciCode48.5%estimated ± 3.8 pp, high confidence
AIME2695.9%estimated ± 1.2 pp, medium confidence
Artificial Analysis Intelligence Index17.9%measured
CritPt1.4%measured
GDPval-AA21.4%estimated ± 6.2 pp, medium confidence
GPQA Diamond89.7%estimated ± 1.8 pp, high confidence
HLE20.4%estimated ± 8.5 pp, medium confidence
IFBench57.7%estimated ± 6.6 pp, medium confidence
AA-LCR55.3%measured
LiveCodeBench v689.0%estimated ± 2.2 pp, medium confidence
AA-Omniscience Accuracy45.8%measured
SciCode47.9%estimated ± 2.9 pp, medium confidence
Terminal-Bench 2.169.9%estimated ± 6.7 pp, medium confidence
BrowseComp64.0%estimated ± 6.2 pp, medium confidence
IFEval91.7%estimated ± 2.8 pp, high confidence
LongBench v253.1%estimated ± 5.5 pp, medium confidence
VITA-Bench28.6%estimated ± 10.7 pp, low confidence
τ²-bench results43.3%measured
AA Coding Index59.3%estimated ± 5.6 pp, medium confidence
AA-MMMU-Pro78.6%measured
APEX-Agents-AA15.5%estimated ± 3.6 pp, medium confidence
BioMysteryBench (human-difficult)20.9%estimated ± 6.8 pp, low confidence
DeepSearchQA73.8%estimated ± 8.2 pp, low confidence
SWE-bench Verified79.3%estimated ± 4.4 pp, high confidence
BFCL v438.7%estimated ± 8.5 pp, low confidence
CyberGym55.3%estimated ± 9.3 pp, medium confidence
JobBench11.4%measured
SWE-bench Pro54.6%estimated ± 2.9 pp, medium confidence
WideResearch72.8%estimated ± 1.2 pp, medium confidence
AA Agentic Index21.1%estimated ± 6.8 pp, medium confidence
MMLU-Pro86.9%estimated ± 4.8 pp, high confidence
AA-IFBench55.1%measured
FailureSensorIQ67.5%estimated ± 1.7 pp, high confidence
GPQA89.5%estimated ± 4.5 pp, high confidence
MMLU90.5%estimated ± 0.9 pp, medium confidence
Vals GPQA Diamond87.9%measured
Vals LiveCodeBench85.6%measured
Vals MMLU-Pro88.6%measured
MMMLU86.8%estimated ± 0.7 pp, medium confidence
MMMU85.6%estimated ± 2.7 pp, high confidence
Chess Puzzles40.0%measured
AA MMLU-Pro88.2%estimated ± 0.7 pp, high confidence
Gert Labs56.6%measured
AA EnterpriseOps-Gym37.6%estimated ± 1.2 pp, low confidence
AA Terminal-Bench 2.180.0%estimated ± 2.5 pp, medium confidence
ApprenticeBench0.0%estimated ± 3.1 pp, low confidence
ARC-AGI-183.4%estimated ± 6.5 pp, medium confidence
ARC-AGI-250.3%estimated ± 13.8 pp, low confidence
CursorBench 3.149.0%estimated ± 5.8 pp, low confidence
CursorBench 3.255.5%estimated ± 5.4 pp, low confidence
EBR-bench11.6%estimated ± 8.1 pp, medium confidence
Terminal-Bench 3.03.1%estimated ± 6.9 pp, low confidence
FrontierCode 1.1 Main30.6%estimated ± 5.2 pp, medium confidence
FrontierMath v2 (Tier 4)17.1%measured
FrontierMath v2 (Tiers 1-3)51.2%measured
FrontierSWE v28.0%estimated ± 9.5 pp, low confidence
Furniture Assembly32.1%estimated ± 7.1 pp, medium confidence
MaxxBench0.0%estimated ± 5.6 pp, low confidence
MazeBench0.0%estimated ± 5.1 pp, low confidence
Mystery Game Puzzles20.0%measured
NYT Connections83.1%measured
OfficeQA Pro50.8%estimated ± 6.8 pp, medium confidence
OSWorld-Verified58.7%estimated ± 3.8 pp, low confidence
ParseBench-174.1%measured
Terminal-Bench 4.19.2%estimated ± 14.9 pp, low confidence
Terminal-Bench Hard45.8%estimated ± 3.9 pp, high confidence
Vals SWE-bench75.0%measured
Terminal-Bench 2.1 (Vals AI)53.9%measured
VulcanBench v384.0%estimated ± 5.6 pp, low confidence
HealthBench Professional (length-adjusted, Artificial Analysis)51.3%estimated ± 4.3 pp, low confidence
HealthBench Professional (raw, Artificial Analysis)53.4%estimated ± 1.1 pp, low confidence
AA ITBench39.2%estimated ± 2.8 pp, high confidence
Chartography (tools)82.5%estimated ± 4.1 pp, low confidence
CursorBench 4.030.8%estimated ± 3.4 pp, medium confidence
DeepSWE56.3%estimated ± 7.2 pp, low confidence
FrontierPhysics0.0%estimated ± 4.6 pp, medium confidence
HLE w/o tools45.3%estimated ± 4.9 pp, high confidence
OSWorld 2.05.7%estimated ± 5.9 pp, low confidence
PostTrainBench v1.119.8%estimated ± 3.7 pp, medium confidence
SpireClimb5.8%estimated ± 7.2 pp, medium confidence
StealthMark30.5%estimated ± 3.5 pp, medium confidence
SWE-bench Multimodal57.0%estimated ± 5.6 pp, low confidence
GDP.pdf16.4%estimated ± 4.1 pp, low confidence
AA Terminal-Bench 4.00.0%estimated ± 13.7 pp, low confidence
ARC-AGI-33.6%estimated ± 8.6 pp, low confidence
ARC-AGI-3 Provider Adapter2.7%estimated ± 7.3 pp, low confidence
SkillsBench official leaderboard (BenchFlow)54.6%measured
BinaryAudit36.8%measured
CompileBench84.4%estimated ± 10.0 pp, low confidence
OTelBench18.8%measured
Vibe Code Bench20.2%measured
FACTS Multimodal49.7%measured
ParseBench-277.7%measured
CharXiv88.9%estimated ± 5.6 pp, low confidence
CharXiv w/o tools87.4%estimated ± 1.8 pp, low confidence
ExploitBench31.5%estimated ± 1.9 pp, low confidence
AI2D_TEST94.0%estimated ± 2.0 pp, medium confidence
C-Eval90.9%estimated ± 1.0 pp, medium confidence
Claw-Eval49.2%measured
CountBench97.3%estimated ± 2.9 pp, medium confidence
DynaMath85.5%estimated ± 3.0 pp, medium confidence
ERQA68.4%estimated ± 3.1 pp, high confidence
HMMT Feb 202689.9%estimated ± 1.5 pp, medium confidence
HMMT Nov 202594.5%estimated ± 1.5 pp, medium confidence
MathVision89.3%estimated ± 3.1 pp, high confidence
MCP Atlas29.2%estimated ± 8.6 pp, medium confidence
MCP-Tasks54.8%estimated ± 6.4 pp, low confidence
MMAnswerBench86.9%estimated ± 2.4 pp, medium confidence
MMLU-Redux96.7%estimated ± 0.8 pp, low confidence
MMMU-Pro79.4%estimated ± 1.5 pp, high confidence
NL2Repo42.5%estimated ± 4.7 pp, high confidence
PostTrainBench v119.0%estimated ± 3.1 pp, medium confidence
RealWorldQA85.3%estimated ± 2.9 pp, medium confidence
ScreenSpot Pro92.2%estimated ± 10.1 pp, low confidence
SuperGPQA91.9%estimated ± 9.3 pp, medium confidence
Terminal-Bench 2.061.7%estimated ± 5.8 pp, medium confidence
Toolathlon46.1%estimated ± 7.2 pp, medium confidence
V*100.0%estimated ± 7.5 pp, medium confidence
Video-MME (w/o subtitle)85.6%estimated ± 1.7 pp, medium confidence
Video-MME (with subtitle)87.4%estimated ± 3.7 pp, medium confidence
AA LiveCodeBench88.9%estimated ± 2.6 pp, high confidence
HealthBench Hard12.3%estimated ± 8.6 pp, low confidence
MedXpertQA (MM)76.3%estimated ± 6.5 pp, low confidence
MMLU-Pro (Arcee)82.4%estimated ± 4.0 pp, medium confidence
React Native Evals68.7%estimated ± 3.0 pp, low confidence
ResearchClawBench12.5%estimated ± 1.3 pp, medium confidence
SWE-Rebench32.8%estimated ± 6.5 pp, low confidence
PostTrainBench17.2%estimated ± 3.7 pp, medium confidence
DAYJOB0.0%measured
DAYJOB Healthcare0.0%measured
BioMysteryBench (human-solvable)81.7%estimated ± 0.6 pp, low confidence
HealthBench (length-adjusted)54.1%estimated ± 3.7 pp, medium confidence
HealthBench Professional56.7%estimated ± 4.2 pp, low confidence
HLE-Verified15.3%estimated ± 4.6 pp, low confidence
LABBench282.3%estimated ± 1.8 pp, low confidence
ScarfBench45.9%estimated ± 12.3 pp, low confidence
BlenderWorld (text-only)70.1%estimated ± 2.8 pp, medium confidence
Cookie-Bench77.1%estimated ± 1.3 pp, medium confidence
STEMBench2.5%estimated ± 0.2 pp, medium confidence
SimpleQA53.5%estimated ± 4.3 pp, low confidence
BabyVision w/ Python90.0%estimated ± 4.0 pp, low confidence
OpenHarmony Bench v1.052.7%estimated ± 3.2 pp, high confidence
SEC-Bench Pro80.8%estimated ± 5.6 pp, medium confidence
MRCRv223.5%estimated ± 11.3 pp, low confidence
MMLongBench-Doc59.9%estimated ± 2.6 pp, low confidence
BioBasicsBench34.1%estimated ± 3.8 pp, medium confidence
MMMU-Pro w/ Python81.5%estimated ± 2.0 pp, high confidence
ExploitBench Internal Port0.4%estimated ± 1.4 pp, medium confidence