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 4.6 vs DeepSeek-V4-Flash

AnthropicvsDeepSeek41 shared benchmarks2516 head-to-head
BenchmarkClaude Sonnet 4.6DeepSeek-V4-Flash
AA Agentic Index61.661.3
AA Intelligence48.451.8
AA-LCR7474.3
AA-Omniscience12.2-14.3
arc_agi_186.584
ARC-AGI-260.446
arena_elo14721439
arena_text_factuality14601434
Artificial Analysis Coding Index6369.1
browsecomp76.273.2
coding_arena_elo15231431
critpt3.116.6
gdpval54.852.9
GDPval-AA (Elo)16761395
GDPval-AA v213811189
GPQA Diamond89.990.8
HLE4938.6
HLE (with tools)46.845.1
IFBench56.679.2
itbenchSre39.831.5
livebench_agentic_coding42.646.8
livebench_coding79.375
livebench_data_analysis7879.3
livebench_instruction_following63.265.5
livebench_language76.179.2
livebench_math8786.8
livebench_reasoning84.886.6
MCP Atlas69.569
OmniScience Accuracy40.940.4
OmniScience Non-Hallucination51.615.6
scicode4749.9
simpleqa_verified2945
SWE-bench Pro58.152.6
SWE-bench Verified79.679
TauBench V3 - Banking34.439.4
Terminal-Bench 2.059.156.9
Terminal-Bench 2.171.282.7
Terminal-Bench Hard59.138.6
toolathlon49.470.3
τ²-Bench Telecom (AA run)97.995.6
τ³-Bench30.522.9

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.