DeepSeek: DeepSeek V3.1

Text input Text output
Author's Description

DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context training process, reaching up to 128K tokens, and uses FP8 microscaling for efficient inference. Users can control the reasoning behaviour with the `reasoning` `enabled` boolean. [Learn more in our docs](https://openrouter.ai/docs/use-cases/reasoning-tokens#enable-reasoning-with-default-config) The model improves tool use, code generation, and reasoning efficiency, achieving performance comparable to DeepSeek-R1 on difficult benchmarks while responding more quickly. It supports structured tool calling, code agents, and search agents, making it suitable for research, coding, and agentic workflows. It succeeds the [DeepSeek V3-0324](/deepseek/deepseek-chat-v3-0324) model and performs well on a variety of tasks.

Key Specifications
Cost
$$$
Context
131K
Parameters
671B (Rumoured)
Released
Aug 21, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Temperature Stop Reasoning Presence Penalty Top Logprobs Frequency Penalty Logprobs Tools Include Reasoning Max Tokens Response Format Tool Choice Top P
Features

This model supports the following features:

Response Format Tools Reasoning
Performance Summary

DeepSeek-V3.1, a 671B parameter hybrid reasoning model, demonstrates a balanced performance profile with exceptional reliability. While its speed performance is moderate, ranking in the 29th percentile, it offers competitive pricing, placing in the 55th percentile. A standout feature is its perfect reliability, achieving a 100% success rate across all benchmarks, indicating consistent and stable operation. The model excels in several key areas. It achieved perfect accuracy in Ethics, making it the most accurate model at its price point and among models of similar speed. It also shows strong performance in General Knowledge (99.5% accuracy) and Instruction Following (70% accuracy, 81st percentile). Its reasoning capabilities are solid at 80% accuracy, and it performs well in Mathematics (90.9%) and Coding (89%). The model effectively handles Hallucinations, with a 98% accuracy in identifying fictional concepts. Its lowest accuracy was in Instruction Following, though still respectable at 70%. DeepSeek-V3.1's ability to support both thinking and non-thinking modes, coupled with its long-context training and efficient inference, positions it as a versatile tool for research, coding, and agentic workflows.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.15
Completion $0.75

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
DeepSeek
DeepSeek | deepseek/deepseek-chat-v3.1 131K $0.15 / 1M tokens $0.75 / 1M tokens
Chutes
Chutes | deepseek/deepseek-chat-v3.1 163K $0.15 / 1M tokens $0.75 / 1M tokens
AtlasCloud
AtlasCloud | deepseek/deepseek-chat-v3.1 131K $0.21 / 1M tokens $0.8 / 1M tokens
Novita
Novita | deepseek/deepseek-chat-v3.1 131K $0.15 / 1M tokens $0.75 / 1M tokens
GMICloud
GMICloud | deepseek/deepseek-chat-v3.1 163K $0.15 / 1M tokens $0.75 / 1M tokens
Fireworks
Fireworks | deepseek/deepseek-chat-v3.1 163K $0.56 / 1M tokens $1.68 / 1M tokens
Parasail
Parasail | deepseek/deepseek-chat-v3.1 163K $0.15 / 1M tokens $0.75 / 1M tokens
Parasail
Parasail | deepseek/deepseek-chat-v3.1 163K $0.15 / 1M tokens $0.75 / 1M tokens
DeepInfra
DeepInfra | deepseek/deepseek-chat-v3.1 163K $0.15 / 1M tokens $0.75 / 1M tokens
DeepInfra
DeepInfra | deepseek/deepseek-chat-v3.1 163K $0.21 / 1M tokens $0.79 / 1M tokens
SambaNova
SambaNova | deepseek/deepseek-chat-v3.1 32K $0.15 / 1M tokens $0.75 / 1M tokens
SiliconFlow
SiliconFlow | deepseek/deepseek-chat-v3.1 163K $0.27 / 1M tokens $1 / 1M tokens
WandB
WandB | deepseek/deepseek-chat-v3.1 161K $0.55 / 1M tokens $1.65 / 1M tokens
SambaNova
SambaNova | deepseek/deepseek-chat-v3.1 131K $0.65 / 1M tokens $1.5 / 1M tokens
Google
Google | deepseek/deepseek-chat-v3.1 163K $0.6 / 1M tokens $1.7 / 1M tokens
Chutes
Chutes | deepseek/deepseek-chat-v3.1 163K $0.2 / 1M tokens $0.8 / 1M tokens
Novita
Novita | deepseek/deepseek-chat-v3.1 131K $0.27 / 1M tokens $1 / 1M tokens
BytePlus
BytePlus | deepseek/deepseek-chat-v3.1 128K $0.15 / 1M tokens $0.75 / 1M tokens
Friendli
Friendli | deepseek/deepseek-chat-v3.1 131K $0.15 / 1M tokens $0.75 / 1M tokens
SambaNova
SambaNova | deepseek/deepseek-chat-v3.1 32K $0.15 / 1M tokens $0.75 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by deepseek