Kimi K2.5

moonshotai/kimi-k2.5

by Moonshot·Text·256K context·Released 2026-01-15·Modified MIT

open-weightagentic

Overview

Moonshot AI's open-weight Kimi K2.5, a large MoE tuned for agentic workflows, long-document reasoning and tool use. Competitive with much pricier hosted models on coding and multi-step tasks, and increasingly popular for autonomous agents. Available from several providers under a permissive license, with a big context window for document-heavy work.

Benchmarks

Standardized scores (0–100), normalized across the catalog.

Intelligence80

Coding82

Instruction following88

Providers

LLMRouteHub routes to the cheapest healthy provider by default.

Provider	Input	Output	Throughput	Uptime
Lambda	$0.66	$2.75	88 tok/s	99.4%
Hyperbolic	$0.71	$2.95	82 tok/s	99.3%
Parasail	$0.85	$3.52	76 tok/s	99.1%
DeepInfra	$0.98	$4.10	70 tok/s	99%

Use this model

curl https://llmroutehub.com/api/v1/chat/completions \
  -H "Authorization: Bearer $LLMROUTEHUB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "moonshotai/kimi-k2.5",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'