DeepSeek

DeepSeek Chat 0324

Type
language model
Context
164K tokens
Max output
8K
Released
24 March 2025
API string
deepseek-v3-0324
Weights
open

Prices by provider

Provider Input, $ per 1M Output, $ per 1M In our data since
submodel $0.2 $0.8 31 Jul 2026
lambda_ai $0.2 $0.6 31 Jul 2026
alibaba-cn $0.25 $1 31 Jul 2026
meganova $0.25 $0.88 31 Jul 2026
DeepInfra $0.25 $0.88 31 Jul 2026
NanoGPT $0.25 $0.7 31 Jul 2026
novita-ai $0.27 $1.12 31 Jul 2026
Novita AI $0.27 $1.12 31 Jul 2026
gmi $0.28 $0.88 31 Jul 2026
Jiekou.AI $0.28 $1.14 31 Jul 2026
Hyperbolic $0.4 $0.4 31 Jul 2026
Nebius Token Factory $0.5 $1.5 31 Jul 2026
Cortecs $0.551 $1.654 31 Jul 2026
Baseten $0.77 $0.77 31 Jul 2026
Fireworks AI $0.9 $0.9 31 Jul 2026
Microsoft Foundry $1.14 $4.56 31 Jul 2026
crusoe $1.5 $1.5 31 Jul 2026
SambaNova $3 $4.5 31 Jul 2026

Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.

The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.

Measurement results

Task set Result Run conditions Measured by
Arena Score in French 1,388.26 1,362–1,414
Arena Score, multi-turn dialogue 1,387.82 1,381–1,395
Arena Score in English 1,386.46 1,382–1,391
Arena Score, overall 1,374.95 1,371–1,379
Arena Score, mathematics 1,373.79 1,364–1,384
Arena Score in Russian 1,371.71 1,361–1,382
Arena Score, programming 1,368.81 1,362–1,376
Arena Score, hard prompts 1,365.7 1,361–1,371
Arena Score, creative writing 1,364.71 1,357–1,373
Arena Score, long queries 1,352.73 1,346–1,359
Arena Score, expert questions 1,351.73 1,339–1,364
Arena Score, instruction following 1,345.68 1,340–1,352
Arena Score in Spanish 1,345.17 1,325–1,366
Creative writing (Lech Mazur’s evaluation) 77 % lechmazur/writing Github repository
MATH, difficulty level five 75.55 % · with a tuned harness Epoch evaluations
GPQA Diamond — graduate-level questions 56.82 % · with a tuned harness Epoch evaluations
Aider Polyglot — code edits in six languages 55.1 % · with a tuned harness Aider LLM Leaderboards
Fiction.LiveBench — holding a long context 50 % Fiction.live leaderboard
Mock AIME 2024–2025 — olympiad problems 37.72 % · with a tuned harness Epoch evaluations
WeirdML — unusual machine learning tasks 36.08 % WeirdML Leaderboard
SimpleBench — trick questions 12.64 % SimpleBench Leaderboard
CritPt — physics problems 0 %