Models
One model today, served from our own GPUs in Hyderabad, India.
Model
DeepSeek: DeepSeek V4.1 Flash
deepseek/deepseek-v4.1-flash also accepts deepseek-ai/DeepSeek-V4.1-Flash
DeepSeek-V4.1-Flash served from India by Gödel Machines. 1M context; thinking on by default (turn it off with chat_template_kwargs.thinking=false); prefix caching. Tool calling. Paid and partner requests are not stored.
| Context | 1M-token context (1,048,576 tokens) |
|---|---|
| Plan maximum output | 235,929 tokens |
| Weights | Native FP8 with FP4 expert weights |
| Hugging Face | deepseek-ai/DeepSeek-V4.1-Flash |
| Served in | India |
| Streaming | Yes |
Features
- Sampling controls
- Tool calling
- Reasoning
- Structured output
Pricing
Compare Free, Student and Paid plans, prices and limits.
| Free tier (this console) | Free, with daily allowances (see plan limits) |
|---|---|
| Student pass | Free, for a university email: 3× the free daily limits and 4 requests at once, on the free tier's shared capacity · prompts stored as for free (see Student pass) |
| Paid keys, India | ₹12 / ₹0.25 / ₹50 per million input / cached input / output tokens, + 18% GST · zero data retention (see paid keys) |
| Paid keys, outside India | Coming soon |
For developers
Send this as model. The docs show full requests.
deepseek/deepseek-v4.1-flashcurl https://api.goedelmachines.com/v1/models