DeepSeek
Frontier

DeepSeek-V3-0324

Frontier

March 2025 refresh of DeepSeek-V3. Small quality bump on coding and reasoning over the original V3 release.

DeepSeek-V3-0324 is a frontier AI model from DeepSeek. It costs $0.2900 per million input tokens and $1.14 per million output tokens (blended $0.5450/M), with a 128,000-token context window.

RatesUSD per million tokens. Blend is 70% input, 30% output, our calculation.

INPUT
$0.2900/M
per million input tokens
OUTPUT
$1.14/M
per million output tokens
BLENDED 70/30
$0.5450/M

+3.8%over 91 days · hover to read

DeepSeek-V3-0324 — blended price

Reconstructed from 91 days of recorded rate-card changes. Prices hold flat between changes because that is what a posted price does — no value here is interpolated.

CONTEXT
128,000
tokens

From OpenRouter's listing, read on 17 Nov 2026.

What it is good at
  • Open weights
  • Slightly stronger code than V3
  • 128K context
  • MoE inference economics
Typical use cases
  • Self-hosted production chat
  • Code generation
  • Migration from earlier V3

Benchmarks

vs. best public score
Graduate-level science questions, "Google-proof". Source: DeepSeek (self-reported), GPQA, read 2026-11-10.
LMArena Elo1396 Elo
Crowd-sourced head-to-head preference Elo rating. Source: LMArena (independent), LMArena Text Arena score, read 2026-11-10.
Hand-curated from each provider's published reports and public leaderboards. Methodology varies across sources, treat as directional rather than authoritative.

Also reported by the provider

Results on benchmarks outside the set we compare across models. They are listed for completeness and are deliberately excluded from the bars above and from every ranking on this site, because a score is only comparable against the same benchmark variant. Why benchmark scores do not compare →

  • MATH-500 (Pass@1)94%self-reported by DeepSeek · 2025-03 · DeepSeek-V3-0324 card, comparison chart (read from the chart image; the card text omits it). Not comparable with the MATH column, so shown here and never ranked.
  • MMLU-Pro (EM)81.2%self-reported by DeepSeek · 2025-03 · DeepSeek-V3-0324 card ("MMLU-Pro: 75.9 -> 81.2"). No full MMLU is published for this release. Not comparable with the MMLU column, so shown here and never ranked.

More from DeepSeek

See all 11 →

Price history

6 recorded changes
Recorded price changes for DeepSeek-V3-0324, newest first
DateRateWasNowChange
8 Sept 2026Output$1.12$1.14+1.8%
8 Sept 2026Input$0.2700$0.2900+7.4%
16 Jul 2026Input$0.2400$0.2700+13%
16 Jul 2026Output$0.9000$1.12+24%
2 Jul 2026Output$0.7700$0.9000+17%
2 Jul 2026Input$0.2000$0.2400+20%

Rates per million tokens, as published by the provider or its router listing. Tracking runs since 3 May; anything earlier is not on record. See every tracked price change.

Frequently asked questions

How much does DeepSeek-V3-0324 cost?

DeepSeek-V3-0324 costs $0.2900 per million input tokens and $1.14 per million output tokens, for a blended reference rate of $0.5450 per million tokens.

What is DeepSeek-V3-0324's context window?

DeepSeek-V3-0324 supports up to 128,000 tokens of context in a single request.

What is DeepSeek-V3-0324 best for?

DeepSeek-V3-0324 is well suited to Open weights, Slightly stronger code than V3 and 128K context.

Who makes DeepSeek-V3-0324?

DeepSeek-V3-0324 is developed and served by DeepSeek. It was released in Mar 2025.

Terms used on this page