Remy Reference/Models/Language/Kimi K2.7 Code

Kimi K2.7 Code

Kimi K2.7 Code, a language model from DeepInfra.

modelOverride: { model: "kimi-k2-7-code" }

MMoonshot·served viaDeepInfra·Released 2026

Pricing

Input: $0.74 / 1M
Output: $3.50 / 1M

Specifications

Context window: 262,164 tokens
Max output: 10,000,000 tokens
Released: 2026
Model string: moonshotai/Kimi-K2.7-Code
Capabilities: CodingAgentic

Parameters

Model options, passed in modelOverride.config.

Parameter	Type	Default	Range / options
`top_p` An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered.	Number	1	0.01–1, step 0.01
`min_p` Float that represents the minimum probability for a token to be considered, relative to the probability of the most likely token. Set to 0 to disable this.	Number	0	0–1, step 0.01
`top_k` Sample from the best k (number of) tokens. 0 means off.	Number	0	0–999, step 1
`presence_penalty` Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics.	Number	0	-2–2, step 0.1
`frequency_penalty` Positive values penalize new tokens based on how many times they appear in the text so far, increasing the model's likelihood to talk about new topics.	Number	0	-2–2, step 0.1
`repetition_penalty` Alternative penalty for repetition, but multiplicative instead of additive (> 1 penalize, < 1 encourage).	Number	1	0.01–5, step 0.01
`stop` A sequence where the API will stop generating further tokens.	Text	—	—

Use this model

await mindstudio.generateText({
  message: "...",
  modelOverride: {
    model: "kimi-k2-7-code",
    config: {
      top_p: 1,
      min_p: 0,
    },
  },
});