benchgap
Other · model

Interfaze Beta benchmark scores

As of 2026-10-07, Interfaze Beta (Other) has measured scores on 5 benchmarks and estimated scores on 35 more.

BenchmarkScoreSource
AA-GPQA Diamond88.5%estimated ± 5.4 pp, medium confidence
AA-HLE35.1%estimated ± 4.0 pp, high confidence
Artificial Analysis Intelligence Index31.5%estimated ± 6.2 pp, medium confidence
CritPt8.2%estimated ± 4.1 pp, high confidence
GPQA Diamond89.9%measured
HLE37.9%estimated ± 8.2 pp, medium confidence
AA-Omniscience Accuracy34.7%estimated ± 8.1 pp, medium confidence
AA-MMMU-Pro69.7%estimated ± 4.9 pp, medium confidence
MMLU-Pro86.3%estimated ± 3.0 pp, high confidence
GPQA89.9%measured
ARC-AGI-188.9%estimated ± 2.7 pp, high confidence
ARC-AGI-255.3%estimated ± 10.1 pp, low confidence
Vals GPQA Diamond89.1%estimated ± 2.0 pp, high confidence
Vals MMLU-Pro86.8%estimated ± 1.2 pp, high confidence
HLE w/o tools35.6%estimated ± 4.4 pp, high confidence
CharXiv70.8%estimated ± 5.4 pp, medium confidence
CharXiv w/o tools74.8%estimated ± 1.8 pp, low confidence
C-Eval93.1%estimated ± 0.9 pp, medium confidence
MathVision74.3%estimated ± 3.6 pp, high confidence
MMLU-Redux94.2%estimated ± 2.1 pp, high confidence
MMMU-Pro71.1%measured
ScreenSpot Pro57.3%estimated ± 10.0 pp, low confidence
SuperGPQA75.4%estimated ± 9.2 pp, low confidence
VideoMMMU84.1%estimated ± 1.0 pp, high confidence
ERQA46.6%estimated ± 6.1 pp, low confidence
HealthBench Hard1.7%estimated ± 5.1 pp, low confidence
MedXpertQA (MM)53.7%estimated ± 2.4 pp, high confidence
MedXpertQA (Text)53.5%estimated ± 4.5 pp, medium confidence
MMLU-Pro (Arcee)86.8%estimated ± 3.4 pp, low confidence
ARC-AGI-30.0%estimated ± 2.1 pp, medium confidence
HealthBench Professional50.0%estimated ± 4.8 pp, low confidence
MMMU46.0%estimated ± 0.4 pp, low confidence
SimpleVQA62.9%estimated ± 8.2 pp, medium confidence
MMMLU90.9%measured
MMLU90.4%estimated ± 3.6 pp, low confidence
MMMU-Pro w/ Python73.1%estimated ± 0.6 pp, high confidence
RefCOCO (avg)82.1%measured
RealWorldQA84.2%estimated ± 4.8 pp, medium confidence
Video-MME (with subtitle)84.8%estimated ± 1.8 pp, low confidence
OmniDocBench 1.52.3%estimated ± 5.9 pp, low confidence