DeepSeek
DeepSeek V3.1
DeepSeek-V3.1 is post-trained on the top of DeepSeek-V3.1-Base, which is built upon the original V3 base checkpoint through a two-phase long context extension approach, following the methodology outlined in the original DeepSeek-V3 report. We have expanded our dataset by collecting additional long documents and substantially extending both training phases. The 32K extension phase has been increased 10-fold to 630B tokens, while the 128K extension phase has been extended by 3.3x to 209B tokens. Additionally, DeepSeek-V3.1 is trained using the UE8M0 FP8 scale data format to ensure compatibility with microscaling data formats.
Our price
input / output per 1M
Official
Not available
USD · Published rate · Updated 2026-10-12
First request
Availablecurl https://api.tokely.me/v1/chat/completions \
-H "Authorization: Bearer $TOKELY_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v3-1",
"messages": [{"role": "user", "content": "Hello!"}],
"max_tokens": 1024
}'Set TOKELY_API_KEY to your dashboard key. Requests require a funded Tokely balance. Base URL: https://api.tokely.me/v1.
Capabilities & limits
- Context window
- 163,840 tokens
- Model max output
- 128,000 tokens
- Cached input price
- $0.478335 per 1M tokens
- Provider
- DeepSeek
- API model ID
deepseek-v3-1- Category
- Text
- Endpoints
POST /v1/chat/completions- Tokely max output
- 128,000 tokens / request, including reasoning tokens
- Default output
- 256 tokens. Set an explicit limit for longer responses.
- Request limits
- 16 MB JSON body (including base64) · 128 messages · 32 function tools
- Rate limit
- 60 requests / minute / account, across its keys
- Supported input
- Text and function tool results
- Model modalities
- text. Tokely accepts text, inline images and PDFs where listed above; audio and video input are not enabled.
The context includes your instructions, conversation history, tool definitions and generated output. Keep the input and requested output within the model’s context window.
API reference
Authenticate with Authorization: Bearer YOUR_TOKELY_API_KEY. Use an OpenAI SDK with the Tokely base URL.
- model
- Required string:
deepseek-v3-1 - messages
- An array of system, developer, user, assistant and tool messages.
- max_tokens
- Optional integer, 1–128,000. Defaults to 256.
- stream
- Streaming has not been verified for this model.
- reasoning_effort
- Optional reasoning effort. Supported values depend on the model; omit to use its default.
- tools
- Function definitions executed by your application. Provider-hosted tools are not enabled.
The response contains choices with message or tool_calls, finish_reason, and usage (prompt_tokens, completion_tokens, total_tokens). Streaming sends chat completion chunks.
- 400 / 415
- Invalid request, parameter, endpoint for this model, or content type.
- 401 / 403
- Invalid, revoked or restricted API key.
- 402
- Insufficient account balance or API key budget.
- 429
- Request limit reached. Retry after the indicated delay.
- 502 / 503 / 504
- Upstream failure, temporary service unavailability or request timeout.
Compare nearby prices
| Model | Input | Output | Context |
|---|---|---|---|
| DeepSeek V3.1DeepSeek · this model | $0.478335 | $1.435005 | 163,840 |
| Gemini 2.5 FlashGoogle | $0.21 | $1.75 | 1,048,576 |
| Gemini 3.5 Flash LiteGoogle | $0.21 | $1.75 | 1,048,576 |
| Mistral Large 3Mistral | $0.53 | $1.58 | 262,144 |
| GPT-3.5 TurboOpenAI | $0.528 | $1.584 | 16,385 |
Frequently asked questions
How much does DeepSeek V3.1 cost?
$0.478335 per million input tokens, $1.435005 per million output tokens and $0.478335 per million cached input tokens. Usage is charged to your Tokely balance.
Do I need a separate DeepSeek account?
Use your Tokely API key and balance. A separate provider key is not required.
How do I switch to this model?
Set model to deepseek-v3-1 and use /v1/chat/completions. Keep your Tokely base URL and API key.
What is the context window?
163,840 tokens, including input and output. The Tokely request body and output limits listed above also apply.
More DeepSeek models & prices
Compare DeepSeek V3.1 with other DeepSeek versions and with the official developer rates.
All DeepSeek models on Tokely · Cheap LLM API — Token Prices & Discounts · Compare AI API gateways · Estimate your token cost