DeepSeek V4.1 Flash
DeepSeek
First token 407 ms typical · 547 tok/s fast end, 305 typical
nw-flash points here
More pricing details
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageCompare prices paid per 1M tokens, output speed and context windows for our OpenAI-compatible API.
The energy pricing discount
One model family, several serving options. Pick the dial that fits each request.
DeepSeek
First token 407 ms typical · 547 tok/s fast end, 305 typical
nw-flash points here
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageZhipuAI
First token 1.00 s typical · 298 tok/s fast end, 178 typical
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageZhipuAI
First token 1.05 s typical · 295 tok/s fast end, 193 typical
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageMoonshotAI
First token 1.54 s typical · 286 tok/s fast end, 125 typical
nw-large points here
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageXiaomi
First token 1.36 s typical · 310 tok/s fast end, 106 typical
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageFirst token 697 ms typical · 146 tok/s fast end, 68 typical
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageQwen
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageQwen
First token 747 ms typical · 161 tok/s fast end, 103 typical
nw-small points here
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageBaseten
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageQwen · 2 model IDs
First token 760 ms typical · 193 tok/s fast end, 171 typical
Price paid and percentages at the $10/kWh pay-as-you-go rate.
Per-token detail on the model pageEach variant uses the same model weights. Reasoning, context cap, priority, or price may change.
Try adjusting your search or filters.
We prioritize what customers ask for. Tell us which model you'd like us to host — email us or come talk to the team in our Discord.