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 Sonnet 4.5 vs DeepSeek-V4-Flash

AnthropicvsDeepSeek37 shared benchmarks730 head-to-head
BenchmarkClaude Sonnet 4.5DeepSeek-V4-Flash
AA Agentic Index50.661.3
AA Intelligence37.451.8
AA-LCR68.374.3
AA-Omniscience-0.1-14.3
arc_agi_163.784
ARC-AGI-213.646
arena_elo14561439
arena_text_factuality14481434
Artificial Analysis Coding Index52.169.1
browsecomp24.173.2
BrowseComp (context management)26.173.2
coding_arena_elo13921431
critpt1.116.6
gdpval40.452.9
GPQA Diamond83.490.8
HLE19.838.6
HLE (with tools)3245.1
IFBench57.379.2
imo_answer_bench65.988.4
LiveCodeBench7191.6
MCP Atlas59.569
MMLU-Pro88.286.4
OmniScience Accuracy32.940.4
OmniScience Non-Hallucination50.915.6
scicode4549.9
simplebench54.346.3
simpleqa_verified23.645
SWE-bench Multilingual6873.3
SWE-bench Pro43.652.6
SWE-bench Verified8279
TauBench V3 - Banking24.539.4
Terminal-Bench 2.05056.9
Terminal-Bench 2.155.882.7
Terminal-Bench Hard35.638.6
toolathlon4170.3
τ²-Bench Telecom (AA run)78.195.6
τ³-Bench1922.9

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.