Bonus: Top up now and we'll double your first deposit — get x2 credits instantly.

DeepSeek V3.1 Terminus: Self-Hosting & Deployment Guide

DeepSeek-V3.1 Terminus is an update to DeepSeek V3.1 that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's…

What is DeepSeek V3.1 Terminus?

DeepSeek-V3.1 Terminus is an update to DeepSeek V3.1 that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's…

Developer: DeepSeek. Released 22 Sep 2025. Context window 163,840 tokens, up to 32,768 output tokens.


How to run DeepSeek V3.1 Terminus yourself

Open weights (mit), 684.5B parameters. Below: how much memory the weights need at each quantization, which hardware fits them, and the commands to serve it.

DeepSeek V3.1 Terminus VRAM requirements

Precision / quantizationWeightsFigureSource
BF16 (full precision)1369.1 GBEstimate684.5B params × 2 bytes
AWQ 4-bit348.9 GBFile sizeadlik/DeepSeek-V3.1-Terminus-AWQ-W4AFP8
MLX 4-bit377.6 GBFile sizemlx-community/DeepSeek-V3.1-Terminus-4bit
GGUF Q8_0713.3 GBFile sizeunsloth/DeepSeek-V3.1-Terminus-GGUF
GGUF Q4_K_M405.4 GBFile sizeunsloth/DeepSeek-V3.1-Terminus-GGUF
GGUF Q2_K245.7 GBFile sizeunsloth/DeepSeek-V3.1-Terminus-GGUF

File size = actual download size on Hugging Face. Estimate = parameter count × bytes per weight. Weights only — the KV cache for long contexts needs extra memory; the fit tables below add 15% headroom as a rule of thumb.

GPU requirements: deploy DeepSeek V3.1 Terminus on your hardware

HardwareBF16 (full precision)AWQ 4-bitMLX 4-bitGGUF Q8_0GGUF Q4_K_MGGUF Q2_K
NVIDIA H200141 GB
NVIDIA H10080 GB
NVIDIA A100 80GB80 GB
NVIDIA L40S48 GB

GPUs needed per variant.

HardwareGGUF Q8_0GGUF Q4_K_MGGUF Q2_K
RTX 509032 GB
RTX 409024 GB
RTX 309024 GB

GGUF via llama.cpp; more cards or CPU offload for larger files.

HardwareGGUF Q8_0GGUF Q4_K_MGGUF Q2_KMLX 4-bit
Mac, 512 GBM3 UltraFits
Mac, 192 GBM2 Ultra
Mac, 128 GBM4 Max / M3 Max
Mac, 64 GBM4 Pro / Max

Assumes ~75% of unified memory is usable by the GPU.

HardwareBF16 (full precision)AWQ 4-bitMLX 4-bitGGUF Q8_0GGUF Q4_K_MGGUF Q2_K
AWS p5.48xlarge8× H100 · 640 GBFitsFitsFitsFits
AWS p4de.24xlarge8× A100 80GB · 640 GBFitsFitsFitsFits
AWS g6e.12xlarge4× L40S · 192 GB
Google Cloud a3-highgpu-8g8× H100 · 640 GBFitsFitsFitsFits
Azure ND H100 v58× H100 · 640 GBFitsFitsFitsFits

Whole-instance GPU memory; public instance specs.

Serve it yourself

pip install -U vllm
vllm serve deepseek-ai/DeepSeek-V3.1-Terminus
pip install 'sglang[all]'
python -m sglang.launch_server --model-path deepseek-ai/DeepSeek-V3.1-Terminus
llama-server -hf unsloth/DeepSeek-V3.1-Terminus-GGUF:Q4_K_M \
  --ctx-size 32768 --n-gpu-layers 99
pip install -U mlx-lm
mlx_lm.generate --model mlx-community/DeepSeek-V3.1-Terminus-4bit --prompt "Hello"

Commands use the official repositories above; vLLM and SGLang expose an OpenAI-compatible endpoint.

Not on LLM.API yet

We do not route this model through our API at the moment. Browse the models you can call today — most workloads have a close match already live.

Providers

Companies that host this model today, with their public list prices. LLM API does not route this model yet.

List price by provider ($ / 1M tokens)

InputOutput
SiliconFlow$0.27 in
$1 out
Novita$0.27 in
$1 out
AtlasCloud$0.3 in
$0.95 out
StreamLake$0.3426 in
$1.0284 out

Provider list prices from OpenRouter's public catalogue.

ProviderInput /MOutput /MCache read /MContextPrecisionUptime (24h)
SiliconFlow$0.27$1163,840fp898.7%
Novita$0.27$1$0.135131,072fp8100.0%
AtlasCloud$0.3$0.95$0.13131,072fp898.4%
StreamLake$0.3426$1.0284128,00099.9%

Source: OpenRouter public catalogue. Last updated 23 Sep 2026.

Not available on LLM API yet

We do not route this model through our API at the moment, so there is no endpoint or code snippet for it yet. Browse the models you can call today — most workloads have a close match already live.

Browse available models

Why Build on LLM.API?

One unified API. Every major model. Built-in reliability, cost control, and observability.

  • Intelligent Model Routing

    Dynamically route each request to the optimal model across providers based on latency, price, and quality—without changing your integration or redeploying code.

    One API, many models
  • Cost-Aware Orchestration

    Automatically balance premium and budget models with per-call controls, caps, and policies so you can ship fast while keeping AI spend predictable and optimized.

    Control spend by design
  • Resilient Fallback Flows

    Define cascading provider and model fallbacks so timeouts, quota limits, or regional outages transparently fail over—maintaining uptime without custom retry logic.

    Fail soft, not hard
  • End-to-End Observability

    Trace every request across providers with logs, metrics, and structured payloads to debug prompts, compare models, and tune performance from a single dashboard.

    See every token
  • Task-Level Abstractions

    Call high-level tasks like chat, generate, extract, or rank instead of vendor-specific APIs, so you can swap models without rewriting business logic.

    Code to tasks, not vendors
  • High-Throughput Batch Jobs

    Run large-scale inference workloads as managed batches with concurrency, retries, and progress tracking built in—perfect for backfills, evaluations, and data processing.

    Crush your backlogs

5 Core Capabilities

  • Refreshed V3.1 checkpoint

    A September update to V3.1 that keeps the same architecture and 128K context while re-tuning output consistency.

  • Cleaner language mixing

    The update targets stray Chinese-English mixing and occasional odd characters reported in V3.1 output.

  • Stronger agent behaviour

    Tuning concentrated on the code agent and search agent paths, the workloads V3.1 was already used for.

  • Hybrid inference retained

    Thinking and non-thinking modes still share one set of weights, chosen by prompt template.

  • Open weights under MIT

    Published openly, so multiple providers host it and the same checkpoint can be self-hosted.

6 Most Valuable Use Cases

  • Drop-in replacement for V3.1 in existing agent pipelines
  • Coding agents where stray language mixing in output caused rework
  • Search-and-summarise agents chaining retrieval with generation
  • Long-document analysis inside a 128K context window
  • Bilingual Chinese and English content workflows
  • Cost-controlled reasoning at open-weight pricing

When to Use — When NOT to Use

Use it if...

  • You already run V3.1 and want its output consistency fixes
  • Your agents mix tool calls with long-context reading
  • You want V3.1 behaviour at a lower published reference price
  • You need open weights you can host yourself

Avoid if...

  • You need multimodal input — this model is text only
  • Your prompts exceed 128,000 tokens
  • You want the newest DeepSeek generation rather than a V3.1 refresh
  • You depend on benchmarks that were only published for the original V3.1

DeepSeek V3.1 Terminus benchmark scores

Intelligence index

This model14
Tracked median12

Scale: 0-100 index points

Reference price per 1M tokens

Input$0.27
Output$1.00

Bars compare published reference list prices, not LLM.API pricing.

Artificial Analysis Intelligence Index14
Median index across all tracked models12
Reference input price$0.27 / 1M tokens
Reference output price$1.00 / 1M tokens

Scores as published by Artificial Analysis and the official model card (source). Reference prices are provider list prices, not LLM.API pricing. Figures with no published value are omitted.

What developers say about DeepSeek V3.1 Terminus

Summarised from publicly published developer write-ups and the model's own documentation. Opinions are the sources’, not LLM.API’s.

  • Hands-on write-ups frame Terminus as a consistency release rather than a capability jump over V3.1.
  • Testers single out fewer language-mixing glitches as the most visible day-to-day difference.
  • Independent scoring puts it level with V3.1 on the intelligence index, at a lower published reference price.

Frequently Asked Questions

  • What is DeepSeek V3.1 Terminus?

    DeepSeek-V3.1 Terminus is an update to DeepSeek V3.1 that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's…

  • Who makes DeepSeek V3.1 Terminus?

    DeepSeek V3.1 Terminus is developed by DeepSeek. It was released on 22 Sep 2025.

  • What is the context window of DeepSeek V3.1 Terminus?

    163,840 tokens, with up to 32,768 output tokens per response.

  • How much does DeepSeek V3.1 Terminus cost?

    The reference list price is $0.27 per 1M input tokens and $1 per 1M output tokens. The cheapest host right now is SiliconFlow at $0.27 / $1 per 1M tokens.

  • Which providers host DeepSeek V3.1 Terminus?

    SiliconFlow, Novita, AtlasCloud, StreamLake.

  • What modalities does DeepSeek V3.1 Terminus support?

    Text input and text output. It supports tool calling, structured outputs, JSON mode, reasoning.

  • Can I use DeepSeek V3.1 Terminus through LLM API?

    Not yet. LLM API does not route this model at the moment. Browse the models page for close alternatives you can call today with one API key.

Get one key to every model

Swap your API key. Keep your code.