AI Rate CardDeepSeek V4 Pro vs Kimi K3: API price and specs
DeepSeek V4 Pro is 64% cheaper than Kimi K3 on input price.
DeepSeek V4 Pro is 61% cheaper than Kimi K3 on output price.
DeepSeek V4 Pro is 63% cheaper than Kimi K3 on a 1M-in / 200k-out workload.
DeepSeek V4 Pro has a 0% larger context than Kimi K3.
| Field | DeepSeek V4 Pro | Kimi K3 |
|---|---|---|
| Lowest input, per 1M | $0.16 | $0.44 +178% |
| Median input, per 1M | $1.50 | $3.00 +99.8% |
| Lowest output, per 1M | $0.63 | $1.61 +154.9% |
| Median output, per 1M | $3.25 | $15.00 +361.5% |
| Lowest unit price | — | — |
| Cost of 1M in + 200k out | $0.28 | $0.76 +167.7% |
| Context | 1M | 1M |
| Max output | 1M | 1M |
| Providers | 44 +26.7% | 60 |
| Regions | APAC, EU, GLOBAL, US | APAC, EU, GLOBAL, US |
| Accepts | file, text | application/pdf, file, image, image/bmp, image/gif, image/heic, image/heif, image/jpeg, image/png, image/webp, pdf, text, video, video/3gpp, video/avi, video/mov, video/mp4, video/mpeg, video/mpg, video/webm, video/wmv, video/x-flv |
| Returns | text | text |
| Type | Text models | Text models |
| Free option | no | yes |
| Cheapest provider | trustedrouter | aihubmix |
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, baseten, cometapi, cortecs, deepinfra, digitalocean, edenai, empiriolabs, fastrouter, huggingface, impossibl, io-net, iteracompute, jiekou, kenari, kilo, meganova, moark, nebius, neuralwatt, novita, ofox, ollama, ollama-cloud, opencode, opencode-go, openrouter, orcarouter, poe, requesty, siliconflow, standardcompute, tensorx, tokengo, trustedrouter, vercel, wafer.ai, zenmux
Frequently asked questions
Which is cheaper, DeepSeek V4 Pro or Kimi K3?
DeepSeek V4 Pro is 64% cheaper than Kimi K3 on input price. DeepSeek V4 Pro is 61% cheaper than Kimi K3 on output price. DeepSeek V4 Pro is 63% cheaper than Kimi K3 on a 1M-in / 200k-out workload.
Which has the larger context, DeepSeek V4 Pro or Kimi K3?
DeepSeek V4 Pro: 1M tokens; Kimi K3: 1M tokens.
Where can I use both DeepSeek V4 Pro and Kimi K3?
aiand, aihubmix, aiml, baseten, cometapi, cortecs, deepinfra, digitalocean, edenai, empiriolabs, fastrouter, huggingface, impossibl, io-net, iteracompute, jiekou, kenari, kilo, meganova, moark, nebius, neuralwatt, novita, ofox, ollama, ollama-cloud, opencode, opencode-go, openrouter, orcarouter, poe, requesty, siliconflow, standardcompute, tensorx, tokengo, trustedrouter, vercel, wafer.ai, zenmux