Kimi K2.5
moonshotai/Kimi-K2.5
Open Source · chat · open-weights
Intelligence
Context
262.1K
Max output
—
Input $/1M
$0.44
Output $/1M
$2.00
Modalities
text · image
Released
01 Jan 2026
Intelligence Index via Artificial Analysis · 0–100, higher is better
License: other · moonshotai/Kimi-K2.5
AI summary
● machine-written
Moonshot AI releases Kimi K2.5, multimodal model with 262K context window
Kimi K2.5 is Moonshot AI's native multimodal model supporting text and image inputs, built on Kimi K2 with continued pretraining over approximately 15T mixed visual and text tokens. It delivers performance in general reasoning, visual coding, and agentic tool-calling with a 262,144 token context window. The model is available as open-weights through multiple API providers starting at $0.44 per million input tokens.
What's new
- Supports visual coding capability with multimodal input (text and image)
- 262,144 token context window enables long-horizon reasoning
- Available from 12 providers with prices ranging $0.44–$0.60 per million input tokens
- Scores 78.9 on GPQA benchmark (77th percentile) and 29.4 on Intelligence (78th percentile)
Best for
Visual coding and image understanding tasksLong-context reasoning with up to 262K tokensAgentic tool-calling and multi-agent orchestrationGeneral reasoning across text and visual domains