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

Qwen3.6 35B A3B vs Qwen3 VL 4B (Reasoning)

AlibabavsAlibaba30 shared benchmarks282 head-to-head
BenchmarkQwen3.6 35B A3BQwen3 VL 4B (Reasoning)
AA Agentic Index21.614.4
AA Intelligence32.17.7
AA-LCR66.723
AA-Omniscience-22.2-68.4
AI2D92.784.9
Artificial Analysis Coding Index41.96.7
critpt0.30
EmbSpatial-Bench0.880.7
gdpval27.813.8
GPQA Diamond8664.1
HallusionBench69.864.1
HLE22.24.6
HMMT 202589.153.1
IFBench64.436.6
LiveCodeBench v680.451.3
LVBench71.453.5
MathVista86.479.5
mmlu_redux93.386
MMLU-Pro85.673.6
MMMU-Pro75.357
MV-Bench74.669.3
OmniScience Accuracy18.812.2
OmniScience Non-Hallucination49.58.2
RealWorldQA85.373.2
RefSpatial-Bench0.645.3
scicode35.817.1
supergpqa64.746.8
Terminal-Bench Hard34.81.5
VideoMMMU83.769.4
τ²-Bench Telecom (AA run)95.315.5

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.