llm.ing

DeepSeek R1

DeepSeekopen weightsOpenRouter ↗Hugging Face ↗Artificial Analysis ↗

Curated entry — models.dev lists DeepSeek only by rolling API aliases.

Benchmarks

BenchmarkCategoryScoreVariantProvenanceSourceObserved
AA Agentic Indexagentic1.1—independentaaSep 7, 2026
1.1—mirroropenrouterSep 8, 2026
AA Coding Indexcoding24.6—mirroropenrouterJul 23, 2026
24.6—independentaaJul 24, 2026
AA Intelligence Indexintelligence11.4—independentaaSep 9, 2026
13.1—mirroropenrouterOct 5, 2026
Design Arena — 3dpreference1084—mirroropenrouterNov 19, 2026
Design Arena — codecategoriespreference1123—mirroropenrouterNov 19, 2026
Design Arena — datavizpreference1182—mirroropenrouterOct 17, 2026
1178—mirroropenrouterNov 18, 2026
Design Arena — gamedevpreference1099—mirroropenrouterOct 17, 2026
1080—mirroropenrouterNov 19, 2026
Design Arena — svgpreference1024—mirroropenrouterNov 19, 2026
Design Arena — uicomponentpreference1089—mirroropenrouterNov 19, 2026
Design Arena — websitepreference1138—mirroropenrouterNov 19, 2026
LMArena Textpreference1373—crowdlmarenaNov 9, 2026
Median output speedspeed160 tok/s—independentaaNov 19, 2026
Median time to first tokenspeed1.02s—independentaaNov 17, 2026

Includes index data from Artificial Analysis.

Providers

Live operational stats per hosting endpoint, via OpenRouter.

ProviderQuantContext$/1M in · outUptime 24hLatencyThroughput
DeepInfrafp4164K$0.5 · $2.1599.92%——
Novitafp8164K$0.7 · $2.599.83%——
SiliconFlowfp8164K$0.5 · $2.1899.27%——
StreamLake—128K$0.571 · $2.2999.27%——