Two of the strongest open models. On Roar AI, DeepSeek-V4-Pro costs $1.56 and Kimi K3 costs $3.705 per million input tokens.
| Feature | DeepSeek-V4-Pro | Kimi K3 |
|---|---|---|
| Made by | DeepSeek | Moonshot AI |
| Released | April 2026 | July 2026 |
| Input price, per 1M tokens | $1.56 | $3.705 |
| Output price, per 1M tokens | $3.12 | $18.525 |
| Cached input, per 1M tokens | $0.234 | $0.351 |
| Batch (usually about half price) | No | No |
| Context window | 1M tokens | 1.05M tokens |
| Longest answer | 384K tokens | — |
| Reads | Text | Text, images, video |
| Thinks before answering | Yes, adjustable | Always on |
| Open weights | Yes: MIT | Yes: Kimi K3 License |
Prices are Roar AI’s published rates, live from our price list. Specs are each lab’s own published figures.
DeepSeek-V4-Pro is the stronger pure coder and has the more permissive licence; Kimi K3 leads on web research, reads images and video, and ranked higher overall at launch.
Take 1,000 requests, each about 3,000 tokens in and 1,000 out: a long prompt and a page of answer.
DeepSeek-V4-Pro is 3.8× cheaper for the same job. Models that think before answering spend extra output tokens doing it, so on hard prompts the gap can be wider than the rates suggest.
On Roar AI both are on the same key and the same monthly invoice. Trying the other one is a one-word change:
client.chat.completions.create(model="deepseek-v4-pro", …) client.chat.completions.create(model="kimi-k3", …)
Send the same prompts to both and compare the answers on your own work. Read the docs.
On Roar AI, DeepSeek-V4-Pro costs $1.56 per million input tokens and $3.12 per million output tokens, and Kimi K3 costs $3.705 per million input tokens and $18.525 per million output tokens. For 1,000 requests of about 3,000 tokens in and 1,000 out, that is $7.80 and $29.64.
Kimi K3 reads up to 1.05M tokens at once, against 1M tokens for DeepSeek-V4-Pro.
Yes. On Roar AI both models answer on the same OpenAI-compatible API, with one key and one monthly bill, so switching between them is a change to the model name.
Create a free account and call both models with one API key.