Models / DeepSeek V4.1 Flash, direct
DeepSeek V4.1 Flash, direct
$0.11 / $0.23 per 1M tokens1M context
DeepSeek V4.1 Flash with nothing in between. Fast and cheap, for chat, short Q&A and latency-sensitive work.
Good for
Chat, short answers, classification, extraction, anything where a second look costs more than a retry.
Not for
Tasks that must be right in one go, such as code edits or strict JSON. Sansi was built for those. Use Sansi →
Price example
A request with 2,000 input tokens and an 800-token reply costs $0.000404.
Listed prices are final: tax included, no top-up fee, no platform fee.
Use it
Only the base_url and the model id change; the rest is the OpenAI API you already use.
curl https://router.xiaojins.com/v1/chat/completions \
-H "Authorization: Bearer $THRICEGATE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-flash",
"messages": [{"role": "user", "content": "Hello"}]
}'Questions
- Is this the official DeepSeek model?
- Yes. Requests go to DeepSeek's official API; we do not swap the model behind the id.
- Why is it 20% above DeepSeek's own price?
- That margin pays for the gateway, one bill across models, and the benchmark we rerun when a new base model ships. If you only want the lowest price, call DeepSeek directly.
- Does it support streaming and tools?
- Yes, through the standard OpenAI chat completions fields.