MoonshotAI: Kimi K2.5
Chatmoonshotai/kimi-k2.5Kimi K2.5 is Moonshot AI's native multimodal model, built on Kimi K2 with continued pretraining over roughly 15T mixed visual and text tokens. It delivers state-of-the-art visual coding capability and a self-directed agent swarm paradigm, with strong general reasoning and agentic tool calling. Capabilities include vision, video input, reasoning, tool use, prompt caching, and web search, so it can read screens, diagrams, and clips and then act on them through tools. Both context and output reach 262K tokens; input is $0.60/M tokens with cache reads at $0.10/M, and web search is billed separately. Released 2026-01-27. Available via OpenAI and Anthropic protocols through Ofox.
Providers
/v1/chat/completionsCode Examples
from openai import OpenAIclient = OpenAI(base_url="https://api.ofox.run/v1",api_key="YOUR_OFOX_API_KEY",)response = client.chat.completions.create(model="moonshotai/kimi-k2.5",messages=[{"role": "user", "content": "Hello!"}],)print(response.choices[0].message.content)
Uptime & Status
Benchmarks
MoonshotAI: Kimi K2.5 scores 1432 in the Overall category of the LMArena text leaderboard (style control), ranking #78 of 374 models based on 8,181 human preference votes (updated 2026-07-12).
| Category | Arena Score | 95% CI | Rank | Votes |
|---|---|---|---|---|
| Overall | 1432 | 1425–1438 | #78 of 374 | 8,181 |
| Hard Prompts | 1461 | 1452–1469 | #64 of 374 | 4,538 |
| Coding | 1504 | 1490–1518 | #44 of 369 | 1,805 |
| Math | 1441 | 1417–1466 | #58 of 362 | 513 |
| Creative Writing | 1390 | 1373–1406 | #100 of 372 | 1,273 |
| Instruction Following | 1435 | 1423–1447 | #60 of 374 | 2,231 |
| Chinese | 1472 | 1434–1509 | #75 of 344 | 242 |
Source: LMArena · CC BY 4.0 · Updated 2026-07-12 · Methodology ↗ · Ranks compare models within each category of the LMArena text leaderboard (style control). Scores come from third-party human preference evaluations, not from OFOX.