---
id: "moonshotai/kimi-k2.5"
also: ["moonshotai/kimi-k2.5-0127", "moonshotai/Kimi-K2.5", "hf/moonshotai/Kimi-K2.5", "requesty/moonshot/kimi-k2.5", "vercel/moonshotai/kimi-k2.5"]
card_usd_per_m: "0.6 in, 3.0 out, at moonshotai/int4"
cheapest_list_usd_per_m: "0.45 in, 2.25 out, at deepinfra/fp4 on openrouter"
cheapest_effective_usd_per_m: "0.1505 in, 3.0 out, at moonshotai/int4 on openrouter"
context: 262144
max_output: 65536
input: ["text", "image"]
output: ["text"]
released: "2026-01-27"
api: ["/api/v1/chat/completions", "/api/v1/completions", "/api/v1/messages", "/api/v1/responses"]
params: ["frequency_penalty", "include_reasoning", "logit_bias", "logprobs", "max_tokens", "min_p", "presence_penalty", "reasoning", "repetition_penalty", "response_format", "seed", "stop", "structured_outputs", "temperature", "tool_choice", "tools", "top_k", "top_logprobs", "top_p"]
endpoints: 22
gateways: "openrouter (10), requesty (6), hf (3), vercel (3)"
no_training: 18
trains: 0
retains_prompts: 2
coding: 46.8000
---
# MoonshotAI: Kimi K2.5

Kimi K2.5 is an open-source, native multimodal agentic model built through continual pretraining on approximately 15 trillion mixed visual and text tokens atop Kimi-K2-Base. It seamlessly integrates vision and language understanding with advanced agentic capabilities, instant and thinking modes, as well as conversational and agentic paradigms.