DeepSeek V3.1 API Pricing 2026

Overview

DeepSeek V3.1 was released on August 21, 2025. Pricing starts at $0.150 per million input tokens and $0.750 per million output tokens. The model supports a context window of up to 33K tokens. API access is available through Deepseek.

Context Window

33K

tokens

Pricing

Input$0.150

Output$0.750

per 1M tokens

Speed

Output-

TTFT-

median latency

Capabilities

Tools

Pricing Comparison

Compare DeepSeek V3.1 with 10 similar models by price.

Current Pricing (per 1M tokens)

11 models

Provider	Model	Input $/M	Output $/M	Context	Actions
DS Deepseek	deepseek-chat-v3.1	$0.150	$0.750	32,768	Try
NO Nousresearch	hermes-4-70b	$0.110	$0.380	131,072	Try
MI Mistral AI	mistral-7b-instruct-v0.1	$0.110	$0.190	2,824	Try
A Allenai	olmo-3-7b-think	$0.120	$0.200	65,536	Try
QW Qwen	qwen-2.5-72b-instruct	$0.120	$0.390	32,768	Try
TC Tencent	hunyuan-a13b-instruct	$0.140	$0.570	131,072	Try
BD Baidu	ernie-4.5-vl-28b-a3b	$0.140	$0.560	30,000	Try
E Essential AI	rnj-1-instruct	$0.150	$0.150	32,768	Try
CO Cohere	command-r-08-2024	$0.150	$0.600	128,000	Try
O OpenAI	gpt-4o-mini-2024-07-18	$0.150	$0.600	128,000	Try
MI Mistral AI	ministral-8b-2512	$0.150	$0.150	262,144	Try

* Some models use tiered pricing based on prompt length. Displayed prices are for prompts ≤ 200k tokens.

Compare Providers

DeepSeek V3.1 is available from multiple providers with different pricing and availability.

10

Providers

$0.200

Cheapest Input/M

$0.790

Cheapest Output/M

Chutes

Best Price

Provider Pricing

10 models

Provider	Input $/M	Output $/M	Uptime
C Chutes Best Price	$0.200	$0.800	1.8%
A AtlasCloud	$0.210	$0.800	98.0%
D DeepInfra	$0.210	$0.790	100.0%
S SiliconFlow	$0.270	$1.000	99.6%
OpenRouter	$0.270	$1.000	-
N Novita	$0.270	$1.000	100.0%
W WandB	$0.550	$1.650	100.0%
F Fireworks	$0.560	$1.680	100.0%
Google	$0.600	$1.700	99.4%
S SambaNova	$0.650	$1.500	92.9%

Provider pricing data sourced from OpenRouter and Helicone

Cost vs. Quality

Compare DeepSeek V3.1's benchmark performance against all models.

X-axis:

Y-axis:

Loading chart...

DeepSeek V3.1

Other models

Detailed Benchmark Scores

Intelligence

27.6 (77th pct)

Coding

27.4 (77th pct)

Math

49.7 (60th pct)

MMLU Pro

83.3 (83th pct)

GPQA

73.5 (78th pct)

LiveCodeBench

57.7 (71th pct)

AIME

-

Aider

-

Benchmark data from Artificial Analysis and HuggingFace Open LLM Leaderboard

Try DeepSeek V3.1

Use our Calculator

Estimate your costs based on expected token usage.

Open Cost Calculator

Try in Playground

Test DeepSeek V3.1 directly in your browser.

Open Playground