Sponsored
LimitDigest
Home / Models / API prices / DeepSeek / DeepSeek V4 Pro

DeepSeek V4 Pro API pricing

Official DeepSeek list price, last checked Oct 29, 2026. Updated Oct 29, 2026.

Quick answer

DeepSeek V4 Pro costs $1.32 per 1M input tokens and $3.96 per 1M output tokens, cached input at $0.044, and a 1M-token context window, per DeepSeek's official pricing page (checked Oct 29, 2026). It's priced in the middle of the pack: number 49 of 77 by blended price ($1.98 per 1M tokens at a 3 to 1 input to output mix). On LMArena's Code Arena it ranks #38 of 50 models we price (relative score 0 out of 100). Cheaper pick: MiMo-V2.6 Flash scores 36 against 0 on the same board at $0.175 blended, 11.3x less.

DeepSeek V4 Pro price per million tokens

DeepSeek V4 Pro price per million tokensUSD
Input$1.32
Cached input$0.044
Output$3.96
Blended, 3 parts input to 1 part output$1.98
Context window1M tokens

Official note from DeepSeek: Peak rate. Off-peak is half price.

What DeepSeek V4 Pro costs for real work

List-price estimates for a month, no discounts. Our math, not a quote: change the token counts in the plan vs API calculator.

WorkloadTokens a monthDeepSeek V4 ProGPT-6.1 TerraDeepSeek V4.1 Flash
Light: a side project or personal assistant2M in, 0.5M out$4.62$4.50$1.20
Medium: a small team or one product feature20M in, 5M out$46.20$45.00$12.00
Heavy: a production app200M in, 50M out$462$450$120
Medium with 80% of input cached20M in (16M cached), 5M out$25.78$29.80$7.30

How DeepSeek V4 Pro ranks on LMArena

Relative score: 100 is the top model we price on that board. Rank is among the models in our price table, not the full leaderboard.

BoardRankScore /100Rating
Coding (Code Arena)#38 of 5001447.2
Chat (Text Arena)#26 of 5101457.8
Creative writing (Text Arena creative writing)#19 of 5111446.5
Roleplay and characters (Text Arena multi-turn)#20 of 5101475
Reasoning and math (Text Arena math)#38 of 5101444.1

Closest rivals to DeepSeek V4 Pro

The models nearest in price, other makers first. Ranks are on Code Arena, where we have them.

ModelIn / outBlendedCoding rankVerdict
GPT-6.1 Terra OpenAI$1 / $5$2#11GPT-6.1 Terra wins: same price, higher score. Compare
Muse Spark 1.3 Meta$1.25 / $4.25$2#14Muse Spark 1.3 wins: same price, higher score. Compare
GLM-5.3 GLM (Z.ai)$1.40 / $4.40$2.15#20GLM-5.3 for quality, DeepSeek V4 Pro for price. Compare
GLM-5.2 GLM (Z.ai)$1.40 / $4.40$2.15#22GLM-5.2 for quality, DeepSeek V4 Pro for price. Compare

Cheaper alternatives to DeepSeek V4 Pro

At least 40% cheaper blended, best LMArena score first.

ModelIn / outBlendedCoding rankVerdict
DeepSeek V4.1 Flash DeepSeek$0.30 / $1.20$0.525#12DeepSeek V4.1 Flash wins on both. Compare
MiMo-V2.6 Flash Xiaomi MiMo$0.14 / $0.28$0.175#17MiMo-V2.6 Flash wins on both. Compare
MiMo-V2.6 Pro Xiaomi MiMo$0.435 / $0.87$0.5437#19MiMo-V2.6 Pro wins on both. Compare

More from DeepSeek

FAQ

How much does the DeepSeek V4 Pro API cost per million tokens?

$1.32 per 1M input tokens and $3.96 per 1M output tokens, plus $0.044 per 1M for cached input, as of Oct 29, 2026. A typical 3 to 1 input to output mix works out to $1.98 per 1M tokens.

Does DeepSeek V4 Pro charge more for long prompts?

DeepSeek doesn't list a long-prompt surcharge for DeepSeek V4 Pro. Its pricing note says: Peak rate. Off-peak is half price.

Is DeepSeek V4 Pro cheaper than GPT-6.1 Terra?

No. A medium month (20M input and 5M output tokens) costs $46.20 on DeepSeek V4 Pro and $45.00 on GPT-6.1 Terra at list price. On Code Arena, DeepSeek V4 Pro ranks #38 and GPT-6.1 Terra #11.

How does DeepSeek V4 Pro rank on LMArena?

On Code Arena it is #38 of 50 models we price, with a relative score of 0 out of 100. It ranks on 5 of 8 boards we track.

What is a cheaper alternative to DeepSeek V4 Pro?

DeepSeek V4.1 Flash costs $0.30 in and $1.20 out per 1M tokens, about 3.8x less blended, and scores 46 on Code Arena against 0 for DeepSeek V4 Pro.

Where does this DeepSeek V4 Pro price come from?

From DeepSeek's official pricing page, which we check every day. This page was last checked Oct 29, 2026.

Sources

Related