ModelsCompareBest forBenchmarksStatusPricingAPI

DeepSeek V3

DeepSeek·Operational·Released 2024-12-26text
Context128k
Max output8.192k
Input / 1M-
Output / 1M-
Cached in-
Benchmarks8

Benchmarks

Cited public sources · rank is exact-benchmark, apples-to-apples

AIME 2025 (MathArena)#21/2325% ↗
GPQA Diamond (Epoch AI run)#71/7756.5% ↗
OTIS Mock AIME 2024-2025 (Epoch AI run)#71/7615.8% ↗
LMArena Elo#56/621332.4 ↗
GPQA Diamond#60/7059.1% ↗
MMLU#4/1488.5% ↗
MMLU-Pro#21/2975.9% ↗
SWE-bench Verified#45/4542% ↗

Compare

vs GPT-6 Astravs GPT-5.6 Solvs GPT-5.6 Terravs GPT-5.6 Lunavs Claude Fable 5.1vs Claude Opus 5vs Claude Sonnet 5vs Claude Haiku 4.5

More from DeepSeek

DeepSeek V4.1 FlashDeepSeek V4 ProDeepSeek V3.2DeepSeek R1

Access DeepSeek V3 and every other model through one endpoint with automatic failover: Respan gateway.