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

DeepSeek-V3.2 vs Qwen3.5 397B A17B

DeepSeekvsAlibaba43 shared benchmarks1033 head-to-head
BenchmarkDeepSeek-V3.2Qwen3.5 397B A17B
AA Agentic Index39.853.3
AA Intelligence32.834.3
AA-LCR70.772.7
AA-Omniscience-22.5-30.8
aime_202695.194.2
apexAgents14.515.3
arena_text_factuality14331445
Artificial Analysis Coding Index44.248.2
browsecomp51.469
BrowseComp (context management)67.678.6
browsecomp_zh6570.3
coding_arena_elo13681399
critpt2.91.7
gdpval18.835.8
GPQA Diamond8489.3
HLE40.829
HLE (with tools)40.848.3
HMMT 202590.292.7
hmmt_feb_202592.594.8
hmmt_feb_202684.187.9
hmmt_nov_202590.292.7
IFBench6178.8
imo_answer_bench78.380.9
LiveCodeBench v683.383.6
longbench_v259.863.2
mcpmark3846.1
mmlu_redux93.794.9
MMLU-Pro8687.8
OmniScience Accuracy3330.8
OmniScience Non-Hallucination17.317.3
scicode3942
Seal-049.546.9
simpleqa_verified27.526
SWE-bench Multilingual70.269.3
SWE-bench Pro15.650.9
SWE-bench Verified73.176.4
TauBench V3 - Banking18.813.4
Terminal-Bench 2.046.452.5
Terminal-Bench 2.146.851.3
Terminal-Bench Hard35.640.9
toolathlon35.238.3
τ²-Bench Telecom (AA run)90.695.6
τ³-Bench69.213.4

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.