DeepSeek V4 Flash 0731 pricing

DeepSeek V4 Flash 0731 from DeepSeek runs on 8 providers. The lowest price is $0.06 input and $0.18 output per 1M tokens, on DeepInfra. On Cloudflare Workers AI it costs 7.3× as much.

Updated

Lab
DeepSeek
Context
1.05M
Input
Text
Providers
8
What each provider chargesBlended price per 1M tokens, 3 parts input to 1 part output. The highest is 7.3× the lowest.
  1. DeepInfra$0.09Lowest
  2. DigitalOcean$0.1491.7×
  3. Baseten$0.1631.8×
  4. CoreWeave$0.1651.8×
  5. Cohere$0.1751.9×
  6. Novita AI$0.4625.1×
  7. Alibaba Cloud$0.5285.9×
  8. Cloudflare Workers AI$0.667.3×

Every price, by provider

DeepSeek V4 Flash 0731 prices by provider, in US dollars per 1 million tokens
ProviderInputOutputCachedCache writeBatch inBatch outContextMax output
DeepInfrafp8$0.06$0.18$0.015–$0.048$0.1441.05M384K
DigitalOcean$0.119$0.238$0.0238–––1.05M944K
Basetenfp8$0.13$0.26$0.028–––1.05M384K
fp32Precision$0.13$0.26$0.028
CoreWeavefp8$0.13$0.27$0.05–––262K236K
Coherefp8$0.14$0.28$0.07–––1.05M384K
Novita AIfp8$0.308$0.924$0.0196–––1.05M393K
Alibaba Cloud$0.352$1.056$0.0352–––1M393K
us-east-1Region$0.374$1.122$0.0374
Cloudflare Workers AI$0.44$1.32$0.014–––1.05M944K

Prices in US dollars. Token prices are per 1 million tokens. abracadabra compiles them from each provider’s listing on OpenRouter, the LiteLLM price list and the providers’ own pricing pages, and refreshes them daily. The dataset is free to use: /api/ai-pricing.

See which tasks your agent finishes.

A price says what a call costs. abracadabra reads every conversation your agent has and shows the topics people bring, the tasks that get done and the turn where the rest stop.

Free to start, no card. Or open the playground (opens in a new tab).