evals.report
BenchmarksLabsCompareRun guidesIn the wild
DeepSeekDeepSeek

DeepSeek R1

DeepSeek · DeepSeek. Released Jan 20, 2025.

DeepSeek R1 is a model from DeepSeek in the DeepSeek family, released Jan 20, 2025. evals.report tracks 29 reported DeepSeek R1 benchmark scores across LiveCodeBench Pro, GPQA Diamond, AIME (OTIS Mock), Humanity's Last Exam, LMArena, WeirdML, Artificial Analysis Intelligence Index, Epoch Capabilities Index, and 21 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

Open29 results

Benchmark results 29

Compare this model
BenchmarkCategoryScoreMetricStatusDate
LiveCodeBench ProCoding1284Codeforces EloOfficialJan 20, 2025Details
GPQA DiamondReasoning69.2%accuracyOfficialJan 20, 2025Details
AIME (OTIS Mock)Reasoning53.3%accuracyOfficialJan 20, 2025Details
Humanity's Last ExamReasoning8.5%accuracyOfficialJan 20, 2025Details
LMArenaChat preference1372source-defined ratingOfficialJan 20, 2025Details
WeirdMLCoding36.5%average accuracyOfficialJan 20, 2025Details
Artificial Analysis Intelligence IndexReasoning27.1IndexUnverifiedJan 20, 2025Details
Epoch Capabilities IndexReasoning142.2IndexOfficialJan 20, 2025Details
Aider PolyglotCoding71.4%% correctOfficialJan 20, 2025Details
MMLU-ProReasoning84.9%accuracyVerifiedJan 20, 2025Details
GAIA: A Benchmark for General AI AssistantsAgents30.30%accuracyUnverifiedJan 20, 2025Details
GDPvalAgents248EloOfficialJan 20, 2025Details
LiveCodeBenchCoding61.7%Pass@1UnverifiedJan 20, 2025Details
SciCodeCoding35.7%accuracyUnverifiedJan 20, 2025Details
IFBenchReasoning38.0%accuracyOfficialJan 20, 2025Details
LongBench v2Reasoning58.3%accuracyOfficialJan 20, 2025Details
Global-MMLUReasoning86.0%accuracyUnverifiedJan 20, 2025Details
Arena-Hard-Auto v2.0Chat preference58.0%% win rateOfficialJan 20, 2025Details
EQ-Bench Creative Writing v3Chat preference1500EloVerifiedJan 20, 2025Details
Design ArenaChat preference1193EloVerifiedJan 20, 2025Details
MASK (Model Alignment between Statements and Knowledge)Other57.32Honesty scoreVerifiedJan 20, 2025Details
BigCodeBenchCoding40.5%calibrated Pass@1VerifiedJan 20, 2025Details
SuperGPQAReasoning61.82%accuracyVerifiedJan 20, 2025Details
IMO-BenchReasoning60.8%accuracyVerifiedJan 20, 2025Details
PutnamBenchReasoning1Problems solvedVerifiedJan 20, 2025Details
Vectara Hallucination LeaderboardOther11.3%Hallucination RateOfficialJan 20, 2025Details
PolyMath: Evaluating Mathematical Reasoning in Multilingual ContextsReasoning47.0Difficulty-Weighted Accuracy (DW-ACC)VerifiedJan 20, 2025Details
Online-Mind2WebAgents25.33%Task success rateVerifiedJan 20, 2025Details
MultiNRCReasoning24.27%accuracyOfficialJan 20, 2025Details