VECTOR WIREAI INTELLIGENCE
NVDA$1,847+3.2%MSFT$512+1.1%GOOGL$199-0.4%META$728+2.7%AMD$184-1.2%TSM$212+0.6%PLTR$98+4.1%AI IDX4,821+1.9%
PKT
SEEDRefresh Models Deals Regulatory Sources

Gemma 4 26B A4B vs Qwen3 VL 4B (Reasoning)

GooglevsAlibaba28 shared benchmarks252 head-to-head
BenchmarkGemma 4 26B A4BQwen3 VL 4B (Reasoning)
AA Agentic Index32.114.4
AA Intelligence26.17.7
AA-LCR61.723
AA-Omniscience-50.8-68.4
AI2D88.384.9
AIME 20259074.5
Artificial Analysis Coding Index39.36.7
critpt00
gdpval25.713.8
GPQA Diamond82.364.1
HallusionBench66.164.1
HLE19.34.6
IFBench72.436.6
ifeval96.482.6
LiveCodeBench79.851.3
LiveCodeBench v677.151.3
mathvision82.460
MathVista79.479.5
mmlu_redux92.786
MMLU-Pro85.273.6
MMMU-Pro73.857
OmniScience Accuracy19.112.2
OmniScience Non-Hallucination13.68.2
RealWorldQA72.273.2
scicode40.317.1
supergpqa61.446.8
Terminal-Bench Hard251.5
τ²-Bench Telecom (AA run)43.615.5

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.