NewHow we catch a seller listing the wrong model, and refund you if one does.
DeepSeek logo

DeepSeek V3.1 API

deepseek/deepseek-v3.1Text

Call DeepSeek V3.1 through MartLLM’s OpenAI-compatible API, from $0.00248/M input and $0.00941/M output, served by third-party providers. Made by DeepSeek. 128,000 token context. No MartLLM fee on requests. Adding credit by card costs 5.5% + $0.50.

Also sold directly by: Amazon Bedrock. Here, sellers serve it through their own provider accounts. Why so cheap? · Where the capacity comes from

How much does the DeepSeek V3.1 API cost?

On November 20, 2026, DeepSeek V3.1 costs $0.00248 per million input tokens and $0.00941 per million output tokens through MartLLM, 99.1% below its list price of $0.27 and $1.00. 10 sellers on the market can serve it. It reads up to 128,000 tokens of context. Every request is listed with its cost in the Activity log. MartLLM adds no fee to these prices, and a request that fails is not billed.

What is the cheapest DeepSeek V3.1 API?

On November 20, 2026, the cheapest DeepSeek V3.1 API we found is MartLLM, at $0.00248 per million input tokens and $0.00941 per million output tokens. That is at least 99% less on both input and output than the closest other provider we checked, Vercel AI Gateway ($0.25 and $0.95). DeepInfra and Requesty are higher still. 10 sellers on the MartLLM market can serve it. MartLLM adds no fee to these prices, and a request that fails is not billed.

Prices as of November 20, 2026, 11:00 UTC. This page is rebuilt every hour.

ProviderInput / 1M tokensOutput / 1M tokensMartLLM is lower by
MartLLM$0.00248$0.00941
Vercel AI Gateway$0.25$0.9599%
DeepInfra$0.25$0.9599%
Requesty$0.30$1.0099.1%

MartLLM: the best seller’s price now, no fee added. Other prices from each provider's public price list: Vercel AI Gateway, DeepInfra, Requesty

DeepSeek V3.1 pricing

You save $0.2675 per 1M input tokens and $0.9906 per 1M output tokens against the official price of $0.27 / $1.00. Why so cheap?

Price now · per 1M tokens
Input
$0.00248
Output
$0.00941
Cached input
$0.00129
% off list
99.1% off
17% of input served from cache
Official price · per 1M tokens
Input
$0.27
Output
$1.00
Cached input
$0.165

Prices as of 11:00 UTC.

Each request is sent to an available third-party provider for this model and retried if the provider fails before it starts answering. No MartLLM fee on requests. Adding credit by card costs 5.5% + $0.50.

How to use DeepSeek V3.1 with the OpenAI SDK

With the OpenAI SDK, set the base URL to MartLLM and call DeepSeek V3.1 with your existing code. Anthropic SDKs work against https://api.martllm.com/v1/messages.

python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.martllm.com/v1",
    api_key="ull_your_key_here",
)

resp = client.chat.completions.create(
    model="deepseek/deepseek-v3.1",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)

Use DeepSeek V3.1 in your apps

By API, DeepSeek V3.1 answers on three addresses with the same MartLLM key and the model id deepseek/deepseek-v3.1: OpenAI chat completions at https://api.martllm.com/v1/chat/completions, the OpenAI Responses API at https://api.martllm.com/v1/responses and the Anthropic Messages API at https://api.martllm.com/v1/messages.

How do I use DeepSeek V3.1 in Claude Code?

Claude Code reaches DeepSeek V3.1 through the Anthropic-compatible address of MartLLM. Set ANTHROPIC_BASE_URL to https://api.martllm.com (with no /v1 at the end), ANTHROPIC_AUTH_TOKEN to your MartLLM key and ANTHROPIC_MODEL to deepseek/deepseek-v3.1, then start claude. Set the opus, sonnet, haiku and subagent model variables to deepseek/deepseek-v3.1 as well: a slot left unset keeps asking for a Claude model at Claude prices. CLAUDE_CODE_MAX_CONTEXT_TOKENS is set to 128,000, the context size of DeepSeek V3.1, because Claude Code assumes 200,000 tokens for a model it does not know.

bash
export ANTHROPIC_BASE_URL=https://api.martllm.com
export ANTHROPIC_AUTH_TOKEN=ull_your_key
export ANTHROPIC_MODEL='deepseek/deepseek-v3.1'
export ANTHROPIC_DEFAULT_OPUS_MODEL='deepseek/deepseek-v3.1'
export ANTHROPIC_DEFAULT_SONNET_MODEL='deepseek/deepseek-v3.1'
export ANTHROPIC_DEFAULT_HAIKU_MODEL='deepseek/deepseek-v3.1'
export CLAUDE_CODE_SUBAGENT_MODEL='deepseek/deepseek-v3.1'
export CLAUDE_CODE_MAX_CONTEXT_TOKENS=128000
claude

Claude Code calls: Anthropic Messages API https://api.martllm.com/v1/messages · Full Claude Code guide

How do I use DeepSeek V3.1 in Codex CLI?

Codex CLI reaches DeepSeek V3.1 through the Responses API of MartLLM. In ~/.codex/config.toml, add a model provider with base_url set to https://api.martllm.com/v1 and wire_api = "responses", put your MartLLM key in the UBERLLM_API_KEY environment variable, and set model to deepseek/deepseek-v3.1. model_context_window is set to 128,000, the context size of DeepSeek V3.1, because Codex assumes 272,000 tokens for models outside the list of OpenAI models it knows.

toml
model = "deepseek/deepseek-v3.1"
model_provider = "uberllm"
model_context_window = 128000
web_search = "disabled"

[model_providers.uberllm]
name = "MartLLM"
base_url = "https://api.martllm.com/v1"
env_key = "UBERLLM_API_KEY"
wire_api = "responses"

Codex CLI calls: OpenAI Responses API https://api.martllm.com/v1/responses · Full Codex CLI guide

How do I use DeepSeek V3.1 in OpenCode?

OpenCode reaches DeepSeek V3.1 through the OpenAI-compatible chat address of MartLLM. In ~/.config/opencode/opencode.json, add a provider named uberllm that uses the @ai-sdk/openai-compatible package with baseURL set to https://api.martllm.com/v1, list deepseek/deepseek-v3.1 under its models, put your MartLLM key in the UBERLLM_API_KEY environment variable, then pick uberllm/deepseek/deepseek-v3.1 with /models. The limit numbers are the context size and the largest reply of DeepSeek V3.1; OpenCode uses them to know when to compact a long session.

json
{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "uberllm": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "MartLLM",
      "options": {
        "baseURL": "https://api.martllm.com/v1",
        "apiKey": "{env:UBERLLM_API_KEY}"
      },
      "models": {
        "deepseek/deepseek-v3.1": {
          "name": "DeepSeek V3.1",
          "limit": {
            "context": 128000,
            "output": 8000
          }
        }
      }
    }
  },
  "model": "uberllm/deepseek/deepseek-v3.1"
}

OpenCode calls: OpenAI chat completions https://api.martllm.com/v1/chat/completions · Full OpenCode guide

How do I use DeepSeek V3.1 in Cursor?

In Cursor Settings → Models, paste your MartLLM key into the OpenAI API Key field, turn on the OpenAI base URL override and enter https://api.martllm.com/v1, turn on Use API Key, then add deepseek/deepseek-v3.1 as a custom model and select it in the chat model picker. This is for chat in Cursor: Tab completion keeps using the models Cursor provides.

OpenAI API Key
ull_your_key
OpenAI base URL override
https://api.martllm.com/v1
Custom model
deepseek/deepseek-v3.1

Cursor calls: OpenAI chat completions https://api.martllm.com/v1/chat/completions · Full Cursor guide

How do I use DeepSeek V3.1 in Cline?

In the settings of Cline, set API Provider to OpenAI Compatible, Base URL to https://api.martllm.com/v1, paste your MartLLM key into OpenAI Compatible API Key and set Model ID to deepseek/deepseek-v3.1. Under Model Configuration, set Context Window Size to 128,000: Cline does not read the limits of a model on its own. In the same place, set Max Output Tokens to 8,000.

API Provider
OpenAI Compatible
Base URL
https://api.martllm.com/v1
OpenAI Compatible API Key
ull_your_key
Model ID
deepseek/deepseek-v3.1
Context Window Size
128000
Max Output Tokens
8000

Cline calls: OpenAI chat completions https://api.martllm.com/v1/chat/completions · Full Cline guide

How do I use DeepSeek V3.1 in SillyTavern?

In SillyTavern, open API Connections, set API to Chat Completion and Chat Completion Source to Custom (OpenAI-compatible), enter https://api.martllm.com/v1 in Custom Endpoint (Base URL), paste your MartLLM key into Custom API Key, click Connect, then type deepseek/deepseek-v3.1 in Enter a Model ID.

API
Chat Completion
Chat Completion Source
Custom (OpenAI-compatible)
Custom Endpoint (Base URL)
https://api.martllm.com/v1
Custom API Key
ull_your_key
Enter a Model ID
deepseek/deepseek-v3.1

SillyTavern calls: OpenAI chat completions https://api.martllm.com/v1/chat/completions · Full SillyTavern guide

How do I use DeepSeek V3.1 in Janitor AI?

In Janitor AI, open a chat, click the gear icon and open API Settings. Choose Proxy and add a configuration with Model Name set to deepseek/deepseek-v3.1, Proxy URL set to https://api.martllm.com/v1/chat/completions and your MartLLM key in API Key. Click Save Settings, then refresh the page.

Config Name
MartLLM
Model Name
deepseek/deepseek-v3.1
Proxy URL
https://api.martllm.com/v1/chat/completions
API Key
ull_your_key

Janitor AI calls: OpenAI chat completions https://api.martllm.com/v1/chat/completions · Full Janitor AI guide

How do I use DeepSeek V3.1 in Open WebUI?

In Open WebUI, go to Settings → Admin → Connections and add an OpenAI API connection with URL set to https://api.martllm.com/v1 and your MartLLM key in API Key. Under Advanced, add deepseek/deepseek-v3.1 to Model IDs, and leave API Type on Chat Completions.

URL
https://api.martllm.com/v1
API Key
ull_your_key
Model IDs
deepseek/deepseek-v3.1
API Type
Chat Completions

Open WebUI calls: OpenAI chat completions https://api.martllm.com/v1/chat/completions · Full Open WebUI guide

DeepSeek V3.1 API FAQ

Can I cap what a DeepSeek V3.1 request costs?

Yes. A request costs what its provider charged; the listed price is $0.068 input and $0.252 output per million tokens, and a request can cost more when the cheaper providers are busy. Send provider.max_price with your most per million tokens: the request only goes to providers at or under it and is never billed above it. Each request is listed with its cost in Activity.

How do I call DeepSeek V3.1 via API?

Use any OpenAI-compatible SDK: set the base URL to https://api.martllm.com/v1, your MartLLM API key, and model "deepseek/deepseek-v3.1". Anthropic SDKs work via https://api.martllm.com/v1/messages.

Is DeepSeek V3.1 OpenAI-compatible on MartLLM?

Yes. Every model on MartLLM, including DeepSeek V3.1, is served through one OpenAI-compatible endpoint. Set the base URL and key in your existing client.

Why is DeepSeek V3.1 cheaper on MartLLM?

MartLLM is a market: sellers offer DeepSeek V3.1 capacity below its list price, and each request goes to a seller that can serve it. A request costs what its seller charged, MartLLM adds no fee to it, and every request is listed with its cost in Activity.

Are there other fees?

MartLLM adds no fee to model prices. A card fee of 5.5% + $0.50 is added when you buy credit. With the 5.5% added, MartLLM's DeepSeek V3.1 price is still lower than every other provider in the table.

Embed this price

Show these prices on your own site. The table reads the current prices from MartLLM, so it stays up to date.

Script: the table sets its own height

<script async src="https://martllm.com/embed.js" data-models="deepseek/deepseek-v3.1"></script>

Plain iframe, for sites that allow no scripts

<iframe src="https://martllm.com/embed/prices?models=deepseek/deepseek-v3.1" title="AI model API prices" width="100%" height="127" loading="lazy" style="border:0;max-width:720px"></iframe>

Price badge for a README (Markdown)

[![DeepSeek V3.1 API price on MartLLM](https://api.martllm.com/api/v1/public/badge/deepseek/deepseek-v3.1.svg)](https://martllm.com/models/deepseek/deepseek-v3.1?utm_source=embed&utm_medium=referral&utm_campaign=price-badge)

Every option is in the embed docs.

Compared with other providers: Cheapest DeepSeek API

More DeepSeek models