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

Llama 3.1 Instruct 405B vs Qwen3.6 35B A3B

MetavsAlibaba20 shared benchmarks218 head-to-head
BenchmarkLlama 3.1 Instruct 405BQwen3.6 35B A3B
AA Agentic Index6.321.6
AA Intelligence8.532.1
AA-LCR24.766.7
AA-Omniscience-17.1-22.2
Artificial Analysis Coding Index14.541.9
C-Eval72.590
critpt00.3
gdpval027.8
GPQA Diamond51.586
HLE4.222.2
IFBench3964.4
mmlu_redux86.293.3
MMLU-Pro73.485.6
OmniScience Accuracy23.218.8
OmniScience Non-Hallucination4949.5
scicode29.935.8
SWE-bench Pro11.249.5
SWE-bench Verified24.573.4
Terminal-Bench Hard6.834.8
τ²-Bench Telecom (AA run)1995.3

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.