Made with these models
The 2 models this integration runs.
deepseek-v4-flash on deepseek
deepseek-v4-flashtext-to-text
deepseek-v4-pro on deepseek
deepseek-v4-protext-to-text
deepseek-v4-flash on deepseek
deepseek-v4-flashtext-to-text
deepseek-v4-pro on deepseek
deepseek-v4-protext-to-text
deepseek-v4-flash on deepseek
deepseek-v4-flashtext-to-text
deepseek-v4-pro on deepseek
deepseek-v4-protext-to-text
deepseek-v4-flash on deepseek
deepseek-v4-flashtext-to-text
deepseek-v4-pro on deepseek
deepseek-v4-protext-to-text
Frontier-grade reasoning and code at a fraction of frontier pricing.
- Frontier-class reasoning at roughly a tenth the price
- v4-flash for throughput, v4-pro for hard reasoning
- Charged at conservative cache-miss rates — never under-charged
- Charged on DeepSeek-reported token usage
- Bring your own key, or run on the platform key
DeepSeek V4 delivers reasoning and code quality in the frontier conversation at roughly a tenth of frontier pricing, which changes what is worth automating. Steps you would not spend a frontier model on — reviewing every generated file, reasoning over each item in a batch — become affordable.
Two models: v4-flash for throughput, v4-pro for hard reasoning. Both drop into an LLM Worker like any other provider.
Charged at DeepSeek's published cache-miss rates. That is deliberately the conservative choice — the API does not tell us whether a cache hit occurred, and we would rather over-state our own cost than under-charge and mis-pay a vendor.
llmreasoningcodecheapdeepseekvalue