benchgap
Google · model

Gemini 3.5 Flash benchmark scores

As of 2026-11-22, Gemini 3.5 Flash (Google) has measured scores on 54 benchmarks and estimated scores on 141 more.

BenchmarkScoreSource
AA-GPQA Diamond92.2%measured
AA-HLE42.7%measured
AA-SciCode53.9%measured
AIME2697.7%estimated ± 1.2 pp, medium confidence
Artificial Analysis Intelligence Index50.2%measured
CritPt13.1%measured
GDPval-AA42.2%measured
GPQA Diamond92.7%measured
HLE40.2%measured
IFBench76.3%measured
AA-LCR69.3%measured
LiveCodeBench v689.9%estimated ± 2.2 pp, low confidence
AA-Omniscience Accuracy51.9%measured
SciCode53.1%measured
Terminal-Bench 2.176.2%measured
BrowseComp90.2%estimated ± 4.2 pp, medium confidence
HLE w/ tools57.0%estimated ± 4.1 pp, low confidence
IFEval93.1%estimated ± 2.8 pp, high confidence
LongBench v258.1%estimated ± 5.5 pp, medium confidence
VITA-Bench49.6%estimated ± 9.9 pp, low confidence
τ²-bench results95.3%measured
AA Coding Index70.1%measured
AA-MMMU-Pro84.3%measured
APEX-Agents41.4%estimated ± 4.9 pp, low confidence
APEX-Agents-AA47.1%measured
BioMysteryBench (human-difficult)45.7%estimated ± 6.8 pp, low confidence
DeepSearchQA88.1%estimated ± 4.5 pp, high confidence
SWE-bench Verified76.8%estimated ± 3.6 pp, high confidence
BFCL v476.0%estimated ± 2.7 pp, medium confidence
CyberGym66.5%estimated ± 5.3 pp, medium confidence
JobBench40.1%estimated ± 7.3 pp, low confidence
SkillsBench59.9%estimated ± 9.1 pp, low confidence
SWE-bench Pro55.1%measured
WideResearch81.4%estimated ± 2.1 pp, medium confidence
AA Agentic Index27.3%measured
MMLU-Pro85.2%estimated ± 4.4 pp, high confidence
AA-IFBench76.3%measured
FailureSensorIQ68.2%estimated ± 1.7 pp, high confidence
GPQA92.2%measured
MMLU90.5%estimated ± 0.9 pp, medium confidence
Vals GPQA Diamond92.7%measured
Vals LiveCodeBench87.6%measured
Vals MMLU-Pro89.5%measured
MMMLU90.0%estimated ± 0.7 pp, medium confidence
MMMU88.6%estimated ± 0.8 pp, medium confidence
Chess Puzzles45.0%measured
AA MMLU-Pro89.1%estimated ± 0.7 pp, medium confidence
Gert Labs48.6%measured
AA-AnalystAgent45.0%measured
AA EnterpriseOps-Gym50.1%measured
AA Harvey LAB v1.085.6%estimated ± 1.2 pp, medium confidence
AA Tau3 Banking39.8%estimated ± 2.0 pp, medium confidence
AA Terminal-Bench 2.184.0%estimated ± 2.5 pp, high confidence
ApprenticeBench25.9%estimated ± 5.0 pp, low confidence
ARC-AGI-192.5%measured
ARC-AGI-272.1%measured
BabaIsBench88.0%measured
CursorBench 3.149.8%measured
CursorBench 3.248.8%measured
EBR-bench4.8%measured
Terminal-Bench 3.018.2%estimated ± 6.9 pp, medium confidence
FrontierCode 1.1 Main32.2%estimated ± 1.6 pp, low confidence
FrontierMath v2 (Tier 4)26.8%measured
FrontierMath v2 (Tiers 1-3)62.8%measured
FrontierSWE v27.1%estimated ± 8.3 pp, medium confidence
Furniture Assembly18.6%estimated ± 7.4 pp, medium confidence
MaxxBench20.4%estimated ± 4.3 pp, low confidence
MazeBench0.0%measured
Mystery Game Puzzles32.0%measured
NYT Connections92.6%measured
OfficeQA Pro67.6%estimated ± 6.8 pp, medium confidence
OSWorld-Verified78.4%measured
ParseBench-173.2%measured
Terminal-Bench 4.113.4%estimated ± 10.9 pp, low confidence
Terminal-Bench Hard56.7%estimated ± 3.9 pp, high confidence
Vals SWE-bench78.8%measured
Terminal-Bench 2.1 (Vals AI)74.2%measured
VulcanBench v385.4%estimated ± 5.6 pp, low confidence
CWE-Bench-AA51.3%estimated ± 8.1 pp, medium confidence
AA Cyber Index17.7%estimated ± 9.2 pp, low confidence
DeepsecBench-AA16.7%estimated ± 8.2 pp, medium confidence
Harvey LAB-AA v1.17.4%estimated ± 4.0 pp, low confidence
HealthBench Professional (length-adjusted, Artificial Analysis)61.1%estimated ± 4.3 pp, medium confidence
HealthBench Professional (raw, Artificial Analysis)59.6%estimated ± 0.8 pp, low confidence
AA ITBench40.3%measured
Bug Hunt Bench19.4%estimated ± 9.0 pp, low confidence
Chartography (tools)85.9%estimated ± 4.1 pp, low confidence
CursorBench 4.033.3%estimated ± 2.0 pp, medium confidence
CWE-bench v157.4%estimated ± 5.1 pp, medium confidence
DeepSWE55.8%estimated ± 6.4 pp, medium confidence
DRACO86.9%estimated ± 0.7 pp, low confidence
FrontierCode 2.014.7%estimated ± 5.4 pp, low confidence
FrontierPhysics2.8%estimated ± 5.0 pp, medium confidence
HLE w/o tools45.8%estimated ± 2.7 pp, medium confidence
OSWorld 2.012.9%estimated ± 3.6 pp, high confidence
OSWorld 2.160.2%estimated ± 3.6 pp, high confidence
OSWorld 2.1 Offline35.8%estimated ± 1.9 pp, low confidence
PostTrainBench v1.123.1%estimated ± 3.7 pp, high confidence
ProgramBench50.6%estimated ± 6.0 pp, low confidence
SpireClimb5.8%estimated ± 7.2 pp, medium confidence
StealthMark35.4%estimated ± 8.9 pp, low confidence
SWE-bench Multimodal57.0%estimated ± 5.6 pp, low confidence
Toolathlon-Verified57.0%estimated ± 2.3 pp, medium confidence
VulcanBench Frontier v458.9%estimated ± 2.3 pp, low confidence
AA AutomationBench53.4%estimated ± 9.9 pp, low confidence
GDP.pdf19.5%estimated ± 4.2 pp, high confidence
AA Terminal-Bench 4.04.3%estimated ± 11.0 pp, low confidence
ARC-AGI-30.0%estimated ± 8.3 pp, low confidence
ARC-AGI-3 Provider Adapter2.9%estimated ± 7.3 pp, medium confidence
cursorBench4144.8%estimated ± 3.7 pp, medium confidence
CWE-bench55.8%estimated ± 4.6 pp, high confidence
MKA35.3%estimated ± 4.6 pp, high confidence
SkillsBench official leaderboard (BenchFlow)48.2%measured
BinaryAudit32.4%estimated ± 12.0 pp, low confidence
CompileBench100.0%estimated ± 10.0 pp, low confidence
OTelBench22.5%estimated ± 6.6 pp, medium confidence
Vibe Code Bench48.7%measured
FACTS Multimodal51.7%estimated ± 4.0 pp, high confidence
Hermes Bench67.0%estimated ± 2.9 pp, high confidence
Hermes Index37.2%estimated ± 3.1 pp, high confidence
ParseBench-275.8%measured
CharXiv84.2%measured
CharXiv w/o tools88.5%estimated ± 1.2 pp, high confidence
ExploitBench52.9%estimated ± 7.2 pp, medium confidence
ExploitGym10.9%estimated ± 2.7 pp, high confidence
AI2D_TEST93.5%estimated ± 0.9 pp, low confidence
C-Eval95.2%estimated ± 0.9 pp, low confidence
Claw-Eval63.7%estimated ± 3.2 pp, high confidence
CountBench100.0%estimated ± 2.9 pp, low confidence
DeepPlanning51.7%estimated ± 13.1 pp, low confidence
DynaMath86.9%estimated ± 0.8 pp, low confidence
ERQA68.8%estimated ± 3.1 pp, medium confidence
HMMT Feb 202693.6%estimated ± 1.5 pp, medium confidence
HMMT Nov 202595.2%estimated ± 1.5 pp, low confidence
MathVision93.5%estimated ± 3.0 pp, medium confidence
MCP Atlas83.6%measured
MCP-Tasks79.7%estimated ± 7.0 pp, low confidence
MLVU (M-Avg)87.0%estimated ± 1.5 pp, high confidence
MMAnswerBench89.8%estimated ± 2.4 pp, low confidence
MMLU-Redux97.4%estimated ± 0.8 pp, low confidence
MMMU-Pro83.6%measured
NL2Repo44.7%estimated ± 4.7 pp, high confidence
PostTrainBench v119.7%estimated ± 3.1 pp, medium confidence
QwenClawBench68.9%estimated ± 2.6 pp, medium confidence
RealWorldQA85.1%estimated ± 1.5 pp, high confidence
ScreenSpot Pro78.0%estimated ± 6.8 pp, medium confidence
SimpleVQA72.9%estimated ± 7.5 pp, medium confidence
SuperGPQA80.0%estimated ± 7.2 pp, medium confidence
Terminal-Bench 2.075.6%estimated ± 5.8 pp, medium confidence
Toolathlon56.5%measured
V*100.0%estimated ± 7.5 pp, low confidence
Video-MME (w/o subtitle)87.3%estimated ± 1.3 pp, medium confidence
Video-MME (with subtitle)88.4%estimated ± 2.9 pp, high confidence
VideoMMMU86.0%estimated ± 1.5 pp, medium confidence
We-Math84.7%estimated ± 7.4 pp, low confidence
AA LiveCodeBench90.9%estimated ± 2.6 pp, medium confidence
HealthBench Hard30.2%estimated ± 5.1 pp, low confidence
LiveCodeBench Pro79.8%estimated ± 6.0 pp, low confidence
MedXpertQA (MM)81.7%estimated ± 2.4 pp, high confidence
MedXpertQA (Text)62.7%estimated ± 4.5 pp, medium confidence
MMLU-Pro (Arcee)87.8%estimated ± 3.4 pp, low confidence
React Native Evals81.2%estimated ± 3.0 pp, medium confidence
ResearchClawBench18.0%measured
SWE-Rebench62.0%estimated ± 6.5 pp, low confidence
PostTrainBench20.6%estimated ± 3.7 pp, high confidence
DAYJOB1.3%estimated ± 3.2 pp, high confidence
DAYJOB Healthcare0.7%estimated ± 4.2 pp, high confidence
Finance Agent v257.9%measured
BioMysteryBench (human-solvable)86.1%estimated ± 0.6 pp, low confidence
FrontierCode 1.1 Extended54.3%estimated ± 3.0 pp, low confidence
HealthBench (raw)51.2%estimated ± 6.0 pp, medium confidence
HealthBench (length-adjusted)56.1%estimated ± 2.5 pp, low confidence
HealthBench Professional55.4%estimated ± 4.2 pp, low confidence
HealthBench Professional (raw)61.6%estimated ± 4.5 pp, high confidence
HLE-Verified49.0%estimated ± 2.0 pp, medium confidence
LABBench276.3%estimated ± 1.7 pp, low confidence
ScarfBench41.2%estimated ± 9.1 pp, low confidence
BlenderWorld (text-only)71.3%estimated ± 2.8 pp, medium confidence
Cookie-Bench80.0%estimated ± 1.3 pp, medium confidence
STEMBench2.6%estimated ± 0.2 pp, medium confidence
SimpleQA34.7%estimated ± 7.8 pp, low confidence
Agents' Last Exam31.2%estimated ± 9.7 pp, medium confidence
BabyVision w/ Python94.9%estimated ± 3.8 pp, low confidence
OpenHarmony Bench v1.058.9%estimated ± 2.4 pp, high confidence
SEC-Bench Pro11.5%estimated ± 7.5 pp, low confidence
BabyVision73.0%estimated ± 9.1 pp, low confidence
MMVU80.5%estimated ± 2.7 pp, low confidence
PerceptionBench62.6%estimated ± 4.9 pp, low confidence
MRCRv277.3%measured
MMLongBench-Doc60.9%estimated ± 2.6 pp, low confidence
OCRBench V272.5%estimated ± 3.3 pp, medium confidence
BioBasicsBench29.3%estimated ± 2.2 pp, low confidence
LifeVal77.0%estimated ± 2.3 pp, low confidence
MMMU-Pro w/ Python86.6%estimated ± 2.0 pp, medium confidence
ExploitBench Internal Port0.9%estimated ± 1.4 pp, medium confidence