# DeepSeek-V3.1-Terminus

> DeepSeek's MIT-licensed hybrid-reasoning V3.1 refresh, tuned for cleaner language and stronger code and search agents.

- Maker: DeepSeek
- Line: DeepSeek V3
- Canonical page: https://ai-tldr.dev/models/deepseek-v3-1-terminus/
- Updated: 2026-06-18

## In short

DeepSeek-V3.1-Terminus is an incremental but practical update to DeepSeek's V3.1 line, released by the Chinese AI lab DeepSeek on September 22, 2025 under an MIT license with open weights on Hugging Face. It has a 128K-token context window and costs $0.27 per million input tokens and $0.95 per million output tokens. You can use it through DeepSeek and OpenRouter.

## Facts

| Fact | Value | Source |
| --- | --- | --- |
| Context window | 128K (max output 32K) | [1] |
| Input price | $0.27 / 1M tokens | [2] |
| Output price | $0.95 / 1M tokens | [2] |
| Licence | MIT · open weights | [1] |
| Released | 22 September 2025 | [3] |
| Parameters | 685B total, 37B active (MoE) | [1] |

## Sources

1. [Hugging Face model card](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus)
2. [OpenRouter — pricing & API](https://openrouter.ai/deepseek/deepseek-v3.1-terminus)
3. [DeepSeek-V3.1 release notes (DeepSeek API docs)](https://api-docs.deepseek.com/news/news250821)

## Overview

DeepSeek-V3.1-Terminus is an incremental but practical update to DeepSeek's V3.1 line, released by the Chinese AI lab DeepSeek on September 22, 2025 under an MIT license with open weights on Hugging Face. Rather than a new architecture, Terminus is a refinement: it keeps V3.1's 685B-parameter Mixture-of-Experts design (about 37B parameters active per token) and its hybrid thinking/non-thinking modes, while fixing real-world rough edges users had reported.

The two headline fixes are language consistency and agent quality. DeepSeek-V3.1-Terminus produces far fewer Chinese/English mix-ups and largely eliminates the random stray-character glitches that occasionally appeared in V3.1, giving more coherent multilingual output. On top of that, its built-in Code Agent and Search Agent were tuned to be more reliable inside tool-using and agentic workflows.

The model is text-only and exposes a 128K-class context window (163,840 tokens) with up to 32,768 output tokens. Because it ships as open weights under MIT, DeepSeek-V3.1-Terminus can be self-hosted or accessed through DeepSeek's own API and a range of third-party providers, making it a low-cost option for agentic coding, retrieval and search tasks.

## Benchmarks

- MMLU-Pro: 85% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- GPQA-Diamond: 80.7% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- Humanity's Last Exam: 21.7% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- LiveCodeBench: 74.9% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- Aider-Polyglot: 76.1% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- SWE-bench Verified: 68.4% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- SWE-bench Multilingual: 57.8% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- Terminal-bench: 36.7% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- BrowseComp: 38.5% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- BrowseComp-zh: 45% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))
- SimpleQA: 96.8% ([source](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus))

## Pricing

- Input: $0.27 / 1M tokens per 1M tokens
- Output: $0.95 / 1M tokens per 1M tokens

Pricing shown is the listed rate for DeepSeek-V3.1-Terminus on OpenRouter; the model also ships as open weights for self-hosting. Third-party and official DeepSeek API rates vary.

Source: https://openrouter.ai/deepseek/deepseek-v3.1-terminus

## Strengths

- Open weights under a permissive MIT license — free to self-host, fine-tune, and deploy commercially
- Hybrid model: one set of weights serves both a fast non-thinking chat mode and a deeper reasoning mode
- Strong agentic coding and tool-use scores (SWE-bench Verified 68.4, Terminal-bench 36.7)
- Improved search/browse agent (BrowseComp 38.5) and very high factual recall (SimpleQA 96.8)
- Cleaner multilingual output — fixes V3.1's language mixing and stray-character issues
- Low API pricing relative to comparable frontier models, plus self-hosting option

## Best for

- Agentic coding assistants and autonomous software-engineering loops (SWE-bench-style tasks)
- Terminal and CLI agents that execute multi-step commands
- Web search and browsing agents that gather and synthesize information
- Multilingual chat and content generation with consistent language output
- Long-context document analysis up to 128K tokens
- Cost-sensitive or privacy-sensitive deployments via self-hosted open weights

## How to access

- [DeepSeek](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus) — model id `deepseek-ai/DeepSeek-V3.1-Terminus`
- [OpenRouter](https://openrouter.ai/deepseek/deepseek-v3.1-terminus) — model id `deepseek/deepseek-v3.1-terminus`

## DeepSeek V3 versions

- [DeepSeek-V3.2](https://ai-tldr.dev/models/deepseek-v3-2/) (1 December 2025)
- [DeepSeek-V3.2-Speciale](https://ai-tldr.dev/models/deepseek-v3-2-speciale/) (1 December 2025)
- [DeepSeek-V3.2-Exp](https://ai-tldr.dev/models/deepseek-v3-2-exp/) (29 September 2025)
- DeepSeek-V3.1-Terminus (22 September 2025) — this page
- [DeepSeek-V3.1](https://ai-tldr.dev/models/deepseek-v3-1/) (21 August 2025)
- [DeepSeek-V3-0324](https://ai-tldr.dev/models/deepseek-v3-0324/) (24 March 2025)
- [DeepSeek-V3](https://ai-tldr.dev/models/deepseek-v3/) (26 December 2024)
- [DeepSeek-V2.5](https://ai-tldr.dev/models/deepseek-v2-5/) (5 September 2024)
- [DeepSeek-V2](https://ai-tldr.dev/models/deepseek-v2/) (May 2024)

## FAQ

### Is DeepSeek-V3.1-Terminus open source?

The weights are released under the permissive MIT license on Hugging Face, so you can download, self-host, fine-tune, and use the model commercially. It is best described as open-weights.

### How is Terminus different from DeepSeek-V3.1?

Terminus keeps the same 685B-parameter MoE architecture and hybrid thinking/non-thinking modes, but fixes language-consistency issues (fewer Chinese/English mix-ups and stray characters) and improves the built-in Code Agent and Search Agent. Its agentic benchmarks — SWE-bench, Terminal-bench, BrowseComp and SimpleQA — improve over V3.1.

### Does DeepSeek-V3.1-Terminus support images or other modalities?

No. DeepSeek-V3.1-Terminus is a text-only model — it accepts and produces text, with no vision, audio, or video input.

### What is the context window and max output?

It supports a 128K-class context window of 163,840 tokens, with a maximum output of 32,768 tokens.

## Links

- [Hugging Face model card](https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus)
- [DeepSeek-V3.1 release notes (DeepSeek API docs)](https://api-docs.deepseek.com/news/news250821)
- [OpenRouter — pricing & API](https://openrouter.ai/deepseek/deepseek-v3.1-terminus)
- [Artificial Analysis — independent evaluation](https://artificialanalysis.ai/models/deepseek-v3-1-terminus-reasoning)

Source: AI/TLDR LLM registry, https://ai-tldr.dev/models/deepseek-v3-1-terminus/