Calculate the cost of using kimi-k3 from Perplexity for your AI applications
Pricing data last updated:
Mode: Responses
Built for enterprise-grade reliability, governance, and scale. Deploy in seconds.
kimi-k3 is a responses model from Perplexity, one of 4 responses models they offer. It is priced at $3.00 per 1M input tokens and $15.00 per 1M output tokens, ranking 36 out of 64 responses models by cost and cheaper than 44% of models in this category. kimi-k3 supports function calling, reasoning, web search — one of only 4 responses models with this combination.
Note: Use the interactive calculator above to estimate costs for your specific usage patterns.
At $3.00 per 1M input tokens and $15.00 per 1M output tokens, kimi-k3 ranks 36 out of 64 responses models by input cost. It is more expensive compared to the median of $2.20 for responses models, and is cheaper than 44% of models in this category.
kimi-k3 is one of 4 Perplexity responses models, one of only 4 responses models combining function calling, reasoning, web search.
| Model | Provider | Input / 1M tokens | Output / 1M tokens | vs kimi-k3 |
|---|---|---|---|---|
| openai.gpt-5.4 | Bedrock Mantle | $2.75 | $16.50 | -8% |
| gpt-5.2-codex | OpenAI | $1.75 | $14.00 | -42% |
| gpt-5.3-codex | OpenAI | $1.75 | $14.00 | -42% |
Similar responses models from other providers
No. kimi-k3 costs $3.00 per 1M input tokens while openai.gpt-5.4 costs $2.75 per 1M input tokens, making openai.gpt-5.4 8% more affordable for input. However, kimi-k3 may offer different capabilities or performance characteristics that justify the price difference.
kimi-k3 input pricing is $3.00 per 1M tokens, which is 36% above the median of $2.20 for responses models. It ranks 36 out of 64 responses models by input cost, making it cheaper than 44% of models in this category. For output, it costs $15.00 per 1M tokens compared to the median of $14.00.
Among Perplexity's 4 responses models, kimi-k3 ranks 4 by input cost. Its combination of function calling, reasoning, web search is shared by only 3 other responses models.
The most comparable responses models to kimi-k3 are: openai.gpt-5.4 from Bedrock Mantle ($2.75/1M input tokens); gpt-5.2-codex from OpenAI ($1.75/1M input tokens); gpt-5.3-codex from OpenAI ($1.75/1M input tokens); codex-mini from Azure ($1.50/1M input tokens). These alternatives were selected based on similar capabilities, pricing, and provider diversity. You can compare any of these models in detail using the Bifrost Model Library.
You can monitor Perplexity service health in real time on the Bifrost Perplexity Status page. It tracks current component status, active incidents, and historical reliability data pulled from the official status page every 60 seconds.
kimi-k3 is priced based on input and output tokens. Use the interactive calculator at the top of this page to estimate costs for your specific workload. Enter your expected input and output tokens volume and the calculator will show the total cost breakdown. For reference, processing 1M input tokens costs $3.00 and generating 1M output tokens costs $15.00.