DeepSeek API model list
DeepSeek’s Models & Pricing page lists two models for its own API: deepseek-flash and deepseek-v4-pro. These are the API IDs; the version names identify the models currently served behind them. This list covers DeepSeek’s API, not every DeepSeek model or third-party provider catalog. Pricing and availability can change; checked November 10, 2026. (api-docs.deepseek.com)
| API model ID | Model version listed by DeepSeek | Documented distinction |
|---|---|---|
deepseek-flash |
DeepSeek-V4.1-Flash-1017 | Supports vision; lower listed token rates than Pro. |
deepseek-v4-pro |
DeepSeek-V4-Pro-0813 | Vision is not supported; listed token rates are higher than Flash. |
Both are listed with a 1M-token context length and a maximum output of 384K tokens. DeepSeek’s table lists JSON output, tool calls, the Responses API, and the Anthropic API for both models. It also lists Chat Prefix Completion (Beta) for both; FIM Completion (Beta) is available in non-thinking mode only. These are documented specifications, not a benchmark ranking. (api-docs.deepseek.com)
Which model should you choose?
- Choose
deepseek-flashif your application needs image input. DeepSeek lists Vision for Flash and says Pro does not support it. Flash also has lower published input and output token rates. (api-docs.deepseek.com) - Consider
deepseek-v4-proif you want the Pro model and do not need vision. The published feature table confirms its listed API features and limits, but does not establish that it is better for a particular task. Test your own workload before choosing based on quality, speed, or reliability. (api-docs.deepseek.com) - For text-only workloads, compare your own results and costs. The documented features and prices do not establish a universal “best” model.
DeepSeek API prices
DeepSeek lists model prices in US dollars per 1 million tokens. Input prices depend on cache status and whether usage falls in peak or off-peak hours; output prices also vary by period. (api-docs.deepseek.com)
| Model ID | Input, cache hit (off-peak / peak) | Input, cache miss (off-peak / peak) | Output (off-peak / peak) |
|---|---|---|---|
deepseek-flash |
$0.003 / $0.006 | $0.15 / $0.30 | $0.60 / $1.20 |
deepseek-v4-pro |
$0.022 / $0.044 | $0.66 / $1.32 | $1.98 / $3.96 |
DeepSeek defines peak hours as 01:00–04:00 and 06:00–10:00 UTC, Monday through Friday, excluding Chinese public holidays. All other hours—including weekends and Chinese public holidays in full—are off-peak. Off-peak rates are half the corresponding peak rates. DeepSeek notes that prices may change, so check its pricing page before budgeting. (api-docs.deepseek.com)
Example: For 1 million cache-miss input tokens plus 100,000 output tokens, the listed peak rates give an estimated $0.42 for Flash ($0.30 + $0.12) or $1.716 for Pro ($1.32 + $0.396). This estimate assumes all tokens are billed at those peak rates and excludes any other charges. (api-docs.deepseek.com)
Verify the current model IDs with GET /models
DeepSeek documents GET https://api.deepseek.com/models as the endpoint for listing currently available models. It returns a JSON model list, and each model’s id is the identifier used in API calls. Returned metadata can include a display name, context window, output limit, modalities, reasoning-effort settings, and API-specific capabilities; some metadata may be omitted when unavailable. (api-docs.deepseek.com)
With your DeepSeek API key set in the DEEPSEEK_API_KEY environment variable, check the list with:
curl https://api.deepseek.com/models \
-H "Authorization: Bearer $DEEPSEEK_API_KEY"
The documented response is JSON with object: "list" and a data array of model objects. Check the returned id before configuring a production request. (api-docs.deepseek.com)
Use the API ID, not an assumed version name. DeepSeek says to use deepseek-flash for the current Flash version, DeepSeek-V4.1-Flash-1017. The legacy IDs deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, but the corresponding models are retired; requests using those IDs are served by DeepSeek-V4.1-Flash-1017 and billed at the Flash price. (api-docs.deepseek.com)
Sources
Models & Pricing | DeepSeek API Docs
Lists Models | DeepSeek API Docs