DeepSeek-V4.1-Flash
deepseek-flashThe official alias currently resolves to DeepSeek-V4.1-Flash-1017. Actual Acosmi capabilities follow the selected interface.
- Input
- ¥1 / million tokens
- Output
- ¥4 / million tokens
- Context length
- 1M
- Max output
- 382K
Find a model by name or ID
Prices follow the selected configuration. Actual charges follow the request quote and bill. Billing guide
deepseek-flashThe official alias currently resolves to DeepSeek-V4.1-Flash-1017. Actual Acosmi capabilities follow the selected interface.
deepseek-v4-proDeepSeek-V4-Pro-0813, a text model. Official prices distinguish cache, input/output and peak periods.
glm-5.2A Z.AI text model with a 1M official context window. Acosmi limits and availability follow the selected interface.
happyhorse-1.0-t2vText-to-video, billed by successful output duration and resolution. The text interface does not generate video.
qwen-image-2.0-proImage generation billed per image, not text Tokens. A text catalog entry does not establish image API availability.
qwen3.7-plusQwen model with input-length pricing tiers. Official prices below are for Bailian Beijing, with the temporary discount identified.
Tiered by input length
qwen4-maxImage input · Structured output · Streaming