benchgap
Google · model

Gemini 2.0 Flash-Lite benchmark scores

As of 2026-10-29, Gemini 2.0 Flash-Lite (Google) has measured scores on 5 benchmarks and estimated scores on 37 more.

BenchmarkScoreSource
AA-GPQA Diamond48.3%estimated ± 6.3 pp, medium confidence
AA-HLE4.4%estimated ± 5.0 pp, high confidence
AA-SciCode9.5%estimated ± 3.9 pp, medium confidence
AIME2628.9%estimated ± 4.6 pp, low confidence
CritPt0.4%estimated ± 4.0 pp, high confidence
GPQA Diamond50.9%estimated ± 1.2 pp, high confidence
HLE2.2%estimated ± 6.9 pp, low confidence
SciCode9.4%estimated ± 3.3 pp, medium confidence
AA Coding Index23.4%estimated ± 5.3 pp, low confidence
AA-MMMU-Pro50.1%estimated ± 3.0 pp, high confidence
SWE-bench Verified12.5%measured
SWE-bench Pro0.0%estimated ± 4.6 pp, medium confidence
MMLU-Pro74.5%estimated ± 4.8 pp, medium confidence
AIME 202437.6%estimated ± 2.6 pp, high confidence
Artificial Analysis Intelligence Index10.5%estimated ± 6.3 pp, medium confidence
BBH82.7%estimated ± 4.5 pp, low confidence
GPQA50.5%measured
MMLU87.6%estimated ± 4.1 pp, medium confidence
AA-Omniscience Accuracy20.6%estimated ± 2.7 pp, medium confidence
MATH-50090.2%estimated ± 2.1 pp, high confidence
MMMLU72.8%estimated ± 1.7 pp, high confidence
MMMU65.1%measured
AIME 202523.8%measured
HLE w/o tools11.6%estimated ± 5.2 pp, low confidence
AI2D_TEST87.9%estimated ± 2.0 pp, high confidence
C-Eval81.9%estimated ± 1.2 pp, medium confidence
CountBench93.8%estimated ± 3.2 pp, high confidence
DynaMath80.7%estimated ± 5.8 pp, low confidence
ERQA47.1%estimated ± 5.2 pp, low confidence
MathVision62.5%estimated ± 3.6 pp, high confidence
MLVU (M-Avg)79.4%estimated ± 2.4 pp, high confidence
MMLU-Redux84.0%estimated ± 1.7 pp, high confidence
MMMU-Pro51.3%estimated ± 2.3 pp, high confidence
RealWorldQA77.5%estimated ± 3.6 pp, high confidence
ScreenSpot Pro53.0%estimated ± 7.3 pp, low confidence
SimpleVQA48.1%estimated ± 7.1 pp, medium confidence
Video-MME (w/o subtitle)68.1%estimated ± 1.3 pp, high confidence
Video-MME (with subtitle)75.0%estimated ± 3.2 pp, high confidence
VideoMMMU56.4%estimated ± 3.8 pp, high confidence
We-Math72.4%estimated ± 7.2 pp, low confidence
SimpleQA16.5%measured
MMVU63.4%estimated ± 3.8 pp, medium confidence