whatmodel.sh / DeepSeek V3.1 Terminus

DeepSeek V3.1 Terminus

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

DeepSeek V3.1 Terminus is an open-weights language model from DeepSeek, released September 2025. Its published size is 671B parameters (37B active per token) and a 164k-token context window. Licence: mit.

Catalogue identity notes

Arena entries 'deepseek-v3.1-terminus-thinking' (text), 'deepseek-v3.1-terminus' (text). Maker repo deepseek-ai/DeepSeek-V3.1-Terminus @19510d6: licence tag mit. Parameters: DeepSeek-V3 architecture: '671B total parameters with 37B activated for each token' (HF safetensors metadata 684.5B incl. MTP). Released 2025-09: month the maker's Hugging Face repo was created (createdAt 2025-09-22). One hybrid checkpoint served as deepseek-reasoner (thinking) and deepseek-chat (non-thinking); distinct from the catalogued deepseek-v3-1.

66.5±1.3#121 of 719 LLMsfrom 11 benchmark results across 2 sources

$ how it compares

The LLMs ranked either side of DeepSeek V3.1 Terminus, out of 828 with a score, then more from DeepSeek.

#modelscorewhereRAMreleased
138GigaChat 3.5 Ultra Reasoning (432B-A28B) Sber67±1local~461GB2026-09
139Qwen3.5 35B-A3B Alibaba67±0api $0.35/M +local~23GB2026-02
140DeepSeek V3.1 Terminus DeepSeek67±1local~230GB2025-09
141DeepSeek V3.2 Exp DeepSeek67±1local~726GB2025-09
142Qwen3 235B-A22B Thinking 2507 Alibaba66±1api $0.97/M +local~76GB2025-07
more from DeepSeek
21DeepSeek V4.1 Flash DeepSeek78±1api $0.26/M +local~537GB2026-09
31DeepSeek V4 Pro (0813) DeepSeek76±1api $1.63/M +local~1010GB2026-08
35DeepSeek V3.2 Speciale DeepSeek76±3comparison only

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #123 of 300 · 32% of its score · 5 tests

Beats o3, Gemini 2.5 Pro, Gemini 3.1 Flash-Lite here, though they rank higher overall. Trails SmolLM3 3B, Jamba 1.5 Large, Nova 2 Pro (Preview) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
SWE-bench Verified lab68.4% ✎▁▁▁▁▁▁▂▄█▆▁▁ beats 71/131hard · avg 24%▇▇▇▇ 9.2%sharp · at its level · small run
BrowseComp-ZH lab45.0% ✎▁▁▁▁▁▃·▁▄▅▄█ beats 10/23hard · avg 16%▇▇▇ 7.1%at its level · small run
SWE-bench Multilingual lab57.8% ✎▁··▁▃▂▅▇█▃▃▁ beats 19/60hard · avg 24%▇▇ 6.0%at its level · small run
BrowseComp lab38.5% ✎▂·▄▁▂▃▆▄█▂▃▄ beats 12/42very hard · avg 14%▇▇ 5.4%at its level · small run
Terminal-Bench (pre-2.0 release) lab36.7% ✎▃·▃···▃▃▅·▅█ beats 6/10very hard · avg 14%▇▇ 4.2%at its level · small run
coding▁▁▁▃▆██▇▆▄▃▁ #103 of 458 · 43% of its score · 4 tests

Beats Grok 4.3, Muse Glimmer 30B, Qwen3.5 Plus here, though they rank higher overall. Trails Llama 3.1 Nemotron Ultra 253B v1, Solar Pro 4, Nemotron Cascade 2 30B-A3B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Aider polyglot (225 exercises, two attempts) lab76.1% ✎▂▁▁▃▆▃▄▆█▃▄▂ beats 63/71very hard · avg 13%▇▇▇▇▇▇▇▇▇ 22%sharp · small run
SWE-bench Verified lab68.4% ✎▁▁▁▁▁▁▂▄█▆▁▁ beats 71/131hard · avg 24%▇▇▇▇ 9.2%sharp · at its level · small run
SWE-bench Multilingual lab57.8% ✎▁··▁▃▂▅▇█▃▃▁ beats 19/60hard · avg 24%▇▇ 6.0%at its level · small run
LiveCodeBench (version unspecified) lab74.9% ✎▁···▁▁▃▅▅█▄▅ beats 34/37medium · avg 40%▇▇ 5.5%at its level · small run
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #91 of 393 * · 15% of its score · 1 test

Beats LongCat 2.0 (1.6T-A48B), Inkling Small (266B-A12B), Qwen3.5 122B-A10B here, though they rank higher overall. Trails Gemma 4 12B, Kimi K2 Instruct (0905), Gemma 4 26B-A4B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
LMArena (text)1417.2 ✎▁▁▁▂▃▃▃▆▆▇█▃ beats 267/379avg 1268▇▇▇▇▇▇ 15%blunt
web search█········█·█ #1 of 3 * · 12% of its score · 2 tests
testresultvs the others in its rundifficultycountswhy
BrowseComp-ZH lab45.0% ✎▁▁▁▁▁▃·▁▄▅▄█ beats 10/23hard · avg 16%▇▇▇ 7.1%at its level · small run
BrowseComp lab38.5% ✎▂·▄▁▂▃▆▄█▂▃▄ beats 12/42very hard · avg 14%▇▇ 5.4%at its level · small run
other tests3 not in a task · 25% of its score
testresultvs the others in its rundifficultycountswhy
GPQA Diamond lab80.7% ✎▁▁▁▁▂▄▅▅███▄ beats 201/310hard · avg 43%▇▇▇▇▇ 12%sharp · small run
Humanity's Last Exam lab21.7% ✎▁▁▁▃▆▆▇██▆▁▃ beats 51/104very hard · avg 4.8%▇▇▇ 6.5%small run
MMLU-Pro lab85.0% ✎▁▁·▁▁▁▃▂▄▅█▁ beats 170/188medium · avg 58%▇▇ 6.0%small run

$ where to run it

llama.cpp · recommended

version

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run DeepSeek V3.1 Terminus · needs ~405.4 GB · from huggingface.co/unsloth
    chat in the terminal
    llama cli -hf unsloth/DeepSeek-V3.1-Terminus-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf unsloth/DeepSeek-V3.1-Terminus-GGUF:Q4_K_M

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run DeepSeek V3.1 Terminus · needs ~405.4 GB · from huggingface.co/unsloth

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/unsloth/DeepSeek-V3.1-Terminus-GGUF@Q4_K_M -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run DeepSeek V3.1 Terminus · needs ~403.1 GB · from ollama.com
    chat in the terminal
    ollama run deepseek-v3.1:671b-terminus-q4_K_M

MLX (Mac)

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run DeepSeek V3.1 Terminus · needs ~377.6 GB, Mac only · from huggingface.co/mlx-community
    chat in the terminal
    mlx_lm.chat --model mlx-community/DeepSeek-V3.1-Terminus-4bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model mlx-community/DeepSeek-V3.1-Terminus-4bit

Weights and documentation: huggingface.co/deepseek-ai ↗

$ how the score was worked out · DeepSeek-V3.1-Terminus model… 84%, LMArena leaderboard dataset 15%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
DeepSeek-V3.1-Terminus model card · lab-reported10▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 84%MIT
LMArena leaderboard dataset1▇▇▇▇▇ 15%CC BY 4.0
prior · pull toward the middle▇ 1%
weighted average= 66.5

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json