Models & pricing
Model policy, honest limits, and commercial launch status.
Current model
darkseek-v4.1 is the configured model identifier. Accepted context length must be certified against the public service policy. A model name or a compatible request format is not a long-context guarantee.
Output and capacity
The native gateway applies finite output, body, response, time, concurrency and rate ceilings. Planned public output ceiling: 16,384 tokens; default output: 4,096. An availability guarantee has not been published.
Pricing is not yet published
No retail price, cache discount or prepaid dollar balance is approved yet. Competitor prices are not DarkSeek prices. Do not deposit money or assume a billing contract from this preview. Quota credits and invoices require an explicit commercial policy.