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

Gemini 3 Pro vs Qwen3 VL 4B (Reasoning)

GooglevsAlibaba29 shared benchmarks281 head-to-head
BenchmarkGemini 3 ProQwen3 VL 4B (Reasoning)
AA Agentic Index5214.4
AA Intelligence40.67.7
AA-LCR7323
AA-Omniscience15.3-68.4
AIME 202510074.5
Artificial Analysis Coding Index46.56.7
critpt9.10
gdpval34.213.8
GPQA Diamond91.964.1
HLE45.84.6
IFBench70.436.6
InfoVQA (val)57.279.5
LiveCodeBench9251.3
LiveCodeBench v690.751.3
LongVideoBench77.757.7
LVBench73.553.5
mathvision86.160
MathVista89.879.5
MMLU-Pro90.173.6
MMMU-Pro8157
OCRBench90.381.6
OmniScience Accuracy55.812.2
OmniScience Non-Hallucination108.2
scicode56.117.1
screenspot_pro_no_tools72.749.2
Terminal-Bench Hard56.91.5
Video-MME88.459.7
VideoMMMU87.669.4
τ²-Bench Telecom (AA run)9815.5

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.