AI Rate CardDeepSeek V4 Flash vs DeepSeek V4 Pro: API price and specs
DeepSeek V4 Flash is 48% cheaper than DeepSeek V4 Pro on input price.
DeepSeek V4 Flash is 74% cheaper than DeepSeek V4 Pro on output price.
DeepSeek V4 Flash is 59% cheaper than DeepSeek V4 Pro on a 1M-in / 200k-out workload.
DeepSeek V4 Pro has a 0% larger context than DeepSeek V4 Flash.
| Field | DeepSeek V4 Flash | DeepSeek V4 Pro |
|---|---|---|
| Lowest input, per 1M | $0.0826 | $0.16 +91.6% |
| Median input, per 1M | $0.15 | $1.50 +901.1% |
| Lowest output, per 1M | $0.17 | $0.63 +283.2% |
| Median output, per 1M | $0.33 | $3.25 +882.4% |
| Lowest unit price | — | — |
| Cost of 1M in + 200k out | $0.12 | $0.28 +146.3% |
| Context | 1M | 1M |
| Max output | 1M | 1M |
| Providers | 36 +18.2% | 44 |
| Regions | APAC, EU, GLOBAL, US | APAC, EU, GLOBAL, US |
| Accepts | file, image, image/gif, image/jpeg, image/png, image/webp, text | file, text |
| Returns | text | text |
| Type | Text models | Text models |
| Free option | yes | no |
| Cheapest provider | standardcompute | trustedrouter |
Your own workload: how many tokens the model gets (input) and gives back (output) - for example, in a month. The "Cost" row of the table above is recalculated for these numbers, for each model.
Computed from current offers each time the page opens; the best value of a row is bold, the percent is how far behind the best.
Providers selling both
aiand, aihubmix, aiml, azure, cometapi, deepinfra, deepseek, digitalocean, edenai, fastrouter, huggingface, impossibl, io-net, jiekou, kenari, kilo, meganova, moark, neuralwatt, ofox, ollama-cloud, opencode-go, openrouter, orcarouter, poe, requesty, siliconflow-cn, standardcompute, tokengo, trustedrouter, vercel, volcengine, zenmux
Frequently asked questions
Which is cheaper, DeepSeek V4 Flash or DeepSeek V4 Pro?
DeepSeek V4 Flash is 48% cheaper than DeepSeek V4 Pro on input price. DeepSeek V4 Flash is 74% cheaper than DeepSeek V4 Pro on output price. DeepSeek V4 Flash is 59% cheaper than DeepSeek V4 Pro on a 1M-in / 200k-out workload.
Which has the larger context, DeepSeek V4 Flash or DeepSeek V4 Pro?
DeepSeek V4 Flash: 1M tokens; DeepSeek V4 Pro: 1M tokens.
Where can I use both DeepSeek V4 Flash and DeepSeek V4 Pro?
aiand, aihubmix, aiml, azure, cometapi, deepinfra, deepseek, digitalocean, edenai, fastrouter, huggingface, impossibl, io-net, jiekou, kenari, kilo, meganova, moark, neuralwatt, ofox, ollama-cloud, opencode-go, openrouter, orcarouter, poe, requesty, siliconflow-cn, standardcompute, tokengo, trustedrouter, vercel, volcengine, zenmux