benchgap
Other · model

Gemma 4 26B A4B benchmark scores

As of 2026-11-22, Gemma 4 26B A4B (Other) has measured scores on 23 benchmarks and estimated scores on 128 more.

BenchmarkScoreSource
AA-GPQA Diamond79.2%measured
AA-HLE19.3%measured
AA-SciCode40.0%measured
Artificial Analysis Intelligence Index0.0%estimated ± 5.6 pp, low confidence
CritPt0.0%measured
GDPval-AA3.2%measured
GPQA Diamond82.4%estimated ± 1.2 pp, high confidence
HLE17.2%measured
IFBench70.4%estimated ± 6.6 pp, medium confidence
AA-LCR65.7%measured
LiveCodeBench v677.1%measured
AA-Omniscience Accuracy18.4%measured
SciCode40.1%estimated ± 3.1 pp, high confidence
BrowseComp64.0%estimated ± 8.8 pp, low confidence
HLE w/ tools35.3%estimated ± 5.3 pp, low confidence
IFEval92.9%estimated ± 2.8 pp, high confidence
LongBench v256.4%estimated ± 5.5 pp, medium confidence
τ²-bench results43.6%measured
AA Coding Index39.3%measured
AA-MMMU-Pro69.2%measured
APEX-Agents2.0%estimated ± 9.4 pp, low confidence
APEX-Agents-AA8.8%estimated ± 9.3 pp, low confidence
BioMysteryBench (human-difficult)0.0%estimated ± 6.8 pp, low confidence
DeepSearchQA43.8%estimated ± 9.6 pp, medium confidence
SWE-bench Verified65.9%estimated ± 6.3 pp, medium confidence
BFCL v460.5%estimated ± 2.7 pp, high confidence
CyberGym8.8%estimated ± 6.0 pp, low confidence
JobBench0.0%estimated ± 8.7 pp, low confidence
SWE-bench Pro32.8%estimated ± 6.0 pp, medium confidence
WideResearch61.2%estimated ± 3.3 pp, medium confidence
AA Agentic Index2.2%estimated ± 3.1 pp, high confidence
MMLU-Pro82.6%measured
AA-IFBench72.4%measured
BBH97.4%estimated ± 2.8 pp, medium confidence
GPQA82.3%measured
MMLU81.6%estimated ± 0.9 pp, high confidence
Vals GPQA Diamond57.5%estimated ± 2.9 pp, medium confidence
Vals LiveCodeBench76.7%estimated ± 5.7 pp, low confidence
Vals MMLU-Pro61.1%estimated ± 1.8 pp, medium confidence
MMMLU83.3%estimated ± 0.7 pp, medium confidence
MMMU80.4%estimated ± 0.8 pp, high confidence
Chess Puzzles23.1%estimated ± 10.8 pp, low confidence
AA MMLU-Pro83.8%estimated ± 0.8 pp, medium confidence
Gert Labs19.6%measured
AA Tau3 Banking5.8%estimated ± 4.1 pp, medium confidence
ApprenticeBench3.1%estimated ± 10.4 pp, low confidence
ARC-AGI-113.7%estimated ± 3.9 pp, medium confidence
ARC-AGI-24.2%estimated ± 10.9 pp, low confidence
CursorBench 3.237.7%estimated ± 5.3 pp, low confidence
EBR-bench14.8%estimated ± 11.3 pp, low confidence
FrontierCode 1.1 Main13.6%estimated ± 4.6 pp, medium confidence
FrontierSWE v25.5%estimated ± 14.2 pp, low confidence
Furniture Assembly33.4%estimated ± 12.4 pp, low confidence
MaxxBench37.6%estimated ± 5.7 pp, low confidence
MazeBench0.0%estimated ± 2.7 pp, medium confidence
Mystery Game Puzzles11.4%estimated ± 13.3 pp, low confidence
NYT Connections73.2%estimated ± 6.3 pp, medium confidence
OfficeQA Pro50.3%estimated ± 6.8 pp, low confidence
OSWorld-Verified57.1%estimated ± 8.4 pp, low confidence
ParseBench-158.5%measured
Vals SWE-bench62.3%estimated ± 5.5 pp, medium confidence
CWE-Bench-AA30.7%estimated ± 8.1 pp, low confidence
AA Cyber Index2.8%estimated ± 12.4 pp, low confidence
DeepsecBench-AA5.7%estimated ± 8.2 pp, low confidence
HealthBench Professional (raw, Artificial Analysis)4.3%estimated ± 2.1 pp, low confidence
AA ITBench35.6%estimated ± 4.6 pp, low confidence
Bug Hunt Bench14.6%estimated ± 10.5 pp, low confidence
Chartography (tools)76.4%estimated ± 4.1 pp, low confidence
CursorBench 4.026.6%estimated ± 4.9 pp, medium confidence
CWE-bench v132.6%estimated ± 5.7 pp, low confidence
DeepSWE31.9%estimated ± 7.5 pp, low confidence
DRACO25.2%estimated ± 2.0 pp, low confidence
FrontierCode 2.026.1%estimated ± 9.3 pp, low confidence
FrontierPhysics1.0%estimated ± 4.7 pp, medium confidence
HLE w/o tools8.7%measured
OSWorld 2.11.0%estimated ± 3.6 pp, medium confidence
PostTrainBench v1.111.0%estimated ± 6.2 pp, low confidence
StealthMark21.2%measured
SWE-bench Multimodal9.2%estimated ± 5.6 pp, low confidence
Toolathlon-Verified54.1%estimated ± 3.0 pp, high confidence
VulcanBench Frontier v469.7%estimated ± 6.8 pp, low confidence
AA AutomationBench37.5%estimated ± 9.9 pp, low confidence
GDP.pdf9.4%estimated ± 4.2 pp, medium confidence
ARC-AGI-30.0%estimated ± 7.2 pp, low confidence
ARC-AGI-3 Provider Adapter2.7%estimated ± 7.3 pp, low confidence
CWE-bench25.2%estimated ± 6.2 pp, low confidence
MKA32.7%estimated ± 4.6 pp, medium confidence
Vibe Code Bench14.2%estimated ± 10.0 pp, low confidence
FACTS Multimodal39.5%measured
Hermes Bench14.7%estimated ± 2.9 pp, medium confidence
Hermes Index14.1%estimated ± 5.7 pp, low confidence
ParseBench-262.6%estimated ± 3.5 pp, high confidence
CharXiv83.2%estimated ± 5.6 pp, low confidence
CharXiv w/o tools78.3%estimated ± 1.2 pp, high confidence
ExploitBench50.2%estimated ± 7.3 pp, low confidence
ExploitGym0.0%estimated ± 8.3 pp, low confidence
AI2D_TEST88.1%estimated ± 2.0 pp, medium confidence
C-Eval87.8%estimated ± 0.9 pp, low confidence
CountBench90.9%estimated ± 2.9 pp, low confidence
DynaMath59.5%estimated ± 3.0 pp, low confidence
ERQA42.7%estimated ± 3.1 pp, medium confidence
MathVision82.6%estimated ± 3.0 pp, high confidence
MCP Atlas56.5%estimated ± 10.7 pp, low confidence
MLVU (M-Avg)84.0%estimated ± 2.2 pp, high confidence
MMLU-Redux91.9%estimated ± 0.6 pp, high confidence
MMMU-Pro73.8%measured
NL2Repo25.4%estimated ± 3.4 pp, medium confidence
QwenClawBench55.9%estimated ± 6.0 pp, low confidence
RealWorldQA82.8%estimated ± 2.8 pp, high confidence
ScreenSpot Pro72.8%estimated ± 9.6 pp, low confidence
SimpleVQA61.3%estimated ± 7.5 pp, medium confidence
SuperGPQA58.4%estimated ± 7.2 pp, medium confidence
Toolathlon27.0%estimated ± 6.3 pp, low confidence
V*39.3%estimated ± 7.5 pp, low confidence
Video-MME (w/o subtitle)82.2%estimated ± 1.3 pp, high confidence
Video-MME (with subtitle)84.9%estimated ± 3.0 pp, high confidence
VideoMMMU82.5%estimated ± 2.4 pp, high confidence
We-Math80.4%estimated ± 7.4 pp, low confidence
AA LiveCodeBench87.3%estimated ± 4.7 pp, low confidence
HealthBench Hard0.0%estimated ± 8.6 pp, low confidence
MedXpertQA (MM)60.3%estimated ± 2.4 pp, high confidence
MMLU-Pro (Arcee)77.1%estimated ± 4.0 pp, low confidence
React Native Evals74.0%estimated ± 3.2 pp, high confidence
ResearchClawBench11.7%estimated ± 2.2 pp, low confidence
SWE-Rebench0.0%estimated ± 8.9 pp, low confidence
PostTrainBench9.9%estimated ± 4.1 pp, low confidence
DAYJOB0.0%estimated ± 3.2 pp, medium confidence
DAYJOB Healthcare0.0%estimated ± 4.7 pp, medium confidence
Finance Agent v255.1%estimated ± 2.4 pp, low confidence
BioMysteryBench (human-solvable)81.7%estimated ± 0.6 pp, low confidence
HealthBench (raw)26.9%estimated ± 6.0 pp, low confidence
HealthBench (length-adjusted)45.7%estimated ± 3.4 pp, low confidence
HealthBench Professional58.1%estimated ± 4.5 pp, low confidence
HealthBench Professional (raw)39.0%estimated ± 4.5 pp, medium confidence
ScarfBench24.4%measured
STEMBench2.3%estimated ± 0.2 pp, low confidence
SimpleQA15.9%estimated ± 4.3 pp, medium confidence
Agents' Last Exam29.6%estimated ± 9.7 pp, low confidence
BabyVision w/ Python88.8%estimated ± 3.8 pp, low confidence
OpenHarmony Bench v1.051.0%estimated ± 2.4 pp, medium confidence
SEC-Bench Pro39.5%estimated ± 5.6 pp, medium confidence
BabyVision50.9%estimated ± 9.1 pp, low confidence
MMVU74.6%estimated ± 2.7 pp, medium confidence
PerceptionBench57.4%estimated ± 4.9 pp, low confidence
MRCRv244.1%measured
MMLongBench-Doc60.7%estimated ± 2.6 pp, low confidence
OCRBench V270.0%estimated ± 3.3 pp, high confidence
BioBasicsBench34.1%estimated ± 3.8 pp, low confidence
LifeVal0.6%estimated ± 5.0 pp, low confidence
MMMU-Pro w/ Python73.0%estimated ± 2.0 pp, high confidence
ExploitBench Internal Port0.0%estimated ± 6.8 pp, low confidence