Verified 2026-11-18 · rates follow the live catalog
5 DeepSeek models are in the catalog, from DeepSeek-V4.1-Flash at $0.0072 for one request to DeepSeek-V4-Pro-0813 at $0.03, a 4.2× spread inside one family. Prices are what each model bills at its most expensive hour of the day. One request is a 20,000-token prompt with a 1,000-token reply and nothing cached, an illustration and not a measurement of your traffic.
| Model | Context | Input / 1M | Output / 1M | Cached input / 1M | One request (20,000 in, 1,000 out) | 1,000 requests |
|---|---|---|---|---|---|---|
DeepSeek-V4.1-Flash | 1.0M | $0.30 | $1.20 | $0.006 | $0.0072 | $7.20 |
DeepSeek-V4.1-Flash-1017 | 1.0M | $0.30 | $1.20 | $0.006 | $0.0072 | $7.20 |
DeepSeek-V4.1-Flash-1017-Uncensored | 1.0M | $0.30 | $1.20 | $0.03 | $0.0072 | $7.20 |
DeepSeek-V4-Flash-0731 | 1.0M | $0.44 | $1.32 | $0.014 | $0.01 | $10.12 |
DeepSeek-V4-Pro-0813 | 1.0M | $1.32 | $3.96 | $0.132 | $0.03 | $30.36 |
DeepSeek-V4.1-Flash and DeepSeek-V4.1-Flash-1017 bill the same: $0.30 in and $1.20 out per million tokens. They are identical on every column above: same rates, same context window, same reasoning tiers.| Model | Image input | Video input | Tool calling | Reasoning tiers | Serving precision |
|---|---|---|---|---|---|
DeepSeek-V4.1-Flash | Yes | No | Yes | low, high, max | FP8 |
DeepSeek-V4.1-Flash-1017 | Yes | No | Yes | low, high, max | FP8 |
DeepSeek-V4.1-Flash-1017-Uncensored | Yes | No | Yes | low, high, max | FP8 |
DeepSeek-V4-Flash-0731 | Yes | No | Yes | low, high, max | FP8 |
DeepSeek-V4-Pro-0813 | No | No | Yes | high, max | FP8 |
These models publish a discount against their list price:
DeepSeek-V4.1-Flash: 50% off during 14:00–24:00 UTC, which brings one request to $0.0036 during the cheapest hours.DeepSeek-V4.1-Flash-1017: 50% off during 14:00–24:00 UTC, which brings one request to $0.0036 during the cheapest hours.DeepSeek-V4-Flash-0731: 50% off during 14:00–24:00 UTC, which brings one request to $0.0051 during the cheapest hours.DeepSeek-V4-Pro-0813: 50% off during 14:00–24:00 UTC, which brings one request to $0.015 during the cheapest hours.Each model's page shows its list price, its rates and its schedule.
Free to call, with the last free day shown (UTC, inclusive): DeepSeek-V4.1-Flash-1017-Free through 31 December 2026. Free models never draw down a plan allowance or credits, work on an account with no credits, and answer only streaming requests. After its last free day a model is no longer served. The free AI API page covers how free models behave at their limits.
Create an account, make a key in the dashboard and send it as a bearer token. Between model families, only the model field changes.
curl https://api.openference.com/v1/chat/completions \
-H "Authorization: Bearer $OPENFERENCE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "DeepSeek-V4.1-Flash-1017-Uncensored",
"messages": [{"role": "user", "content": "Say hello"}]}'
Send DeepSeek-V4.1-Flash-1017-Uncensored exactly as written. OpenAI-format clients use https://api.openference.com/v1; Claude Code and other Anthropic-format clients use https://api.openference.com. The OpenAI-compatible API guide covers every endpoint.
See also the cheap AI API guide, the model catalog and pricing.
Across the 5 billed DeepSeek models, input runs $0.30 to $1.32 and output $1.20 to $3.96 per million tokens at the most expensive hour of the day. DeepSeek-V4.1-Flash is the least expensive per request.
https://api.openference.com/v1 for OpenAI-format clients and https://api.openference.com for Anthropic-format clients.
Some do: DeepSeek-V4.1-Flash, DeepSeek-V4.1-Flash-1017, DeepSeek-V4.1-Flash-1017-Uncensored and DeepSeek-V4-Flash-0731 accept image input, and the other one is text-only.
Yes, for now: DeepSeek-V4.1-Flash-1017-Free. Free models are a limited-time offer and answer only streaming requests.
Model catalog · Pricing · Comparisons · Cost guides · Integration guides · FAQ
Guides: GLM coding plan alternative · Free AI API · LLM gateway · LLM proxy · Cheap AI API · OpenAI-compatible API
Model APIs: GLM API · DeepSeek API · Kimi API · Qwen API · MiniMax API