Both models accept text and support tool calls, JSON mode, and structured output. DeepSeek V4 Flash 0731 has the larger context window and lower published token rates. Each model page lists its limits and sample request. Neither model accepts image input.
Query the live catalog
GET /v1/models is public and returns the models currently available, their
capabilities, context limits, and published rates:
id exactly as shown in inference requests. Model IDs are
case-sensitive and use namespace/model format.
Inference usage is deducted from your available credits at the rates above.
See Pricing for welcome credits and how to add more credits.