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 Fable 5 vs DeepSeek-V3.2

AnthropicvsDeepSeek32 shared benchmarks311 head-to-head
BenchmarkClaude Fable 5DeepSeek-V3.2
AA Agentic Index56.639.8
AA Intelligence62.132.8
AA-LCR76.770.7
AA-Omniscience43.3-22.5
arc_agi_190.557
ARC-AGI-276.84
arena_text_factuality14861433
Artificial Analysis Coding Index76.544.2
browsecomp8851.4
coding_arena_elo16261368
critpt28.62.9
deepsearchqa_f194.260.9
gdpval61.918.8
GPQA Diamond92.684
HLE64.540.8
HLE (with tools)64.540.8
IFBench63.561
MCP Atlas84.762.2
OmniScience Accuracy65.333
OmniScience Non-Hallucination36.417.3
PaperBench88.847.1
scicode60.239
simpleqa_verified68.327.5
SWE-bench Multilingual86.670.2
SWE-bench Pro8015.6
SWE-bench Verified9573.1
TauBench V3 - Banking38.118.8
Terminal-Bench 2.18846.8
Terminal-Bench Hard62.935.6
toolathlon61.735.2
τ²-Bench Telecom (AA run)98.590.6
τ³-Bench26.869.2

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.