VISION & IMAGE

Kimi K2.5

Kimi K2.5 is a vision-language model from Moonshot AI, suited to chat, reasoning and coding workloads. It runs with a 256K-token context window.

Specifications

Provider
Moonshot AI
Type
Vision-language model
Category
Vision & image
Context window
256K tokens
License
other
API endpoint
/v1/chat/completions
Model ID
Kimi-K2.5

About 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.

  • transformers
  • safetensors
  • kimi_k25
  • feature-extraction
  • compressed-tensors
  • image-text-to-text
  • conversational
  • custom_code

Hugging Face · 392,940 downloads · 2,902 likes

Kimi K2.5 API example

Chat completions through the OpenAI-compatible Denizen Blu One API. Point any OpenAI SDK at api.denizenblu.com/v1, set model to Kimi-K2.5 and send the request.

curl https://api.denizenblu.com/v1/chat/completions \
  -H "Authorization: Bearer $DENIZEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Kimi-K2.5",
    "messages": [{ "role": "user", "content": "Hello, blue world." }]
  }'

Need keys, streaming or error codes? See the API docs, or try it first in the playground.

Included in every plan

Kimi K2.5 is part of the flat-rate catalog, so you never pay per token to use it. Plans differ by usage limits, concurrency and support, not by whether this model is available.

  • Explorer$11 a year
  • Creator$31 a year
  • Professional$51 a year
  • Ultimate$101 a year
  • EnterpriseContact sales

Compare plans

Browse the full catalog