AI Rate Card

DeepSeek-V3 vs Kimi K2.5: API price and specs

DeepSeek-V3 is 44% cheaper than Kimi K2.5 on input price.

DeepSeek-V3 is 41% cheaper than Kimi K2.5 on output price.

DeepSeek-V3 is 43% cheaper than Kimi K2.5 on a 1M-in / 200k-out workload.

Kimi K2.5 has a 59.8% larger context than DeepSeek-V3.

FieldDeepSeek-V3Kimi K2.5
Lowest input, per 1M$0.25$0.45 +80%
Median input, per 1M$0.32$0.60 +87.5%
Lowest output, per 1M$0.71$1.20 +68.5%
Median output, per 1M$0.94$3.00 +219.5%
Lowest unit price——
Cost of 1M in + 200k out$0.39$0.69 +75.8%
Context164K +37.4%262K
Max output164K +37.4%262K
Providers14 +63.2%38
RegionsEU, USEU, GLOBAL, US
Acceptstextimage, image/bmp, image/gif, image/heic, image/heif, image/jpeg, image/png, image/webp, text, video, video/3gpp, video/avi, video/mov, video/mp4, video/mpeg, video/mpg, video/webm, video/wmv, video/x-flv
Returnstexttext
TypeText modelsText models
Free optionnono
Cheapest providersiliconflowdeepinfra

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

aihubmix, aiml, amazon-bedrock, deepinfra, edenai, fastrouter, helicone, huggingface, novita, poe, requesty, siliconflow, trustedrouter

Frequently asked questions

Which is cheaper, DeepSeek-V3 or Kimi K2.5?

DeepSeek-V3 is 44% cheaper than Kimi K2.5 on input price. DeepSeek-V3 is 41% cheaper than Kimi K2.5 on output price. DeepSeek-V3 is 43% cheaper than Kimi K2.5 on a 1M-in / 200k-out workload.

Which has the larger context, DeepSeek-V3 or Kimi K2.5?

DeepSeek-V3: 164K tokens; Kimi K2.5: 262K tokens.

Where can I use both DeepSeek-V3 and Kimi K2.5?

aihubmix, aiml, amazon-bedrock, deepinfra, edenai, fastrouter, helicone, huggingface, novita, poe, requesty, siliconflow, trustedrouter