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

Claude Opus 4.5 vs Qwen3 VL 4B (Reasoning)

AnthropicvsAlibaba33 shared benchmarks293 head-to-head
BenchmarkClaude Opus 4.5Qwen3 VL 4B (Reasoning)
AA Agentic Index59.614.4
AA Intelligence41.97.7
AA-LCR7623
AA-Omniscience14-68.4
AIME 202592.874.5
Artificial Analysis Coding Index47.86.7
critpt4.60
ERQA46.847.3
gdpval47.313.8
GPQA Diamond8764.1
HLE30.84.6
IFBench5836.6
InfoVQA (val)76.979.5
LiveCodeBench8751.3
LiveCodeBench v684.851.3
LongVideoBench67.257.7
mathvision77.160
MathVista80.279.5
mmlu_redux95.686
MMLU-Pro9073.6
MMMU-Pro7457
MMStar73.273.2
MV-Bench67.269.3
OCRBench86.581.6
OmniScience Accuracy46.612.2
OmniScience Non-Hallucination398.2
OSWorld-Verified66.331.4
RealWorldQA7773.2
scicode5017.1
supergpqa70.646.8
Terminal-Bench Hard471.5
VideoMMMU84.469.4
τ²-Bench Telecom (AA run)89.515.5

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.