Skip to content

Kimi K2.5

ModelOpen sourceActive
by Moonshot AI · family “kimi-k”

Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm. Built on Kimi K2 with continued pretraining over approximately 15T mixed... (description from the OpenRouter listing)

in: textin: imageout: textReasoningTool useStructured output
Epoch Capabilities Index
148.1 #58 of 268
90% CI 146.2 – 149.3
Listing ↗
Released
27 Jan 2026
Context window
262K tokens
Max output
236K tokens
Input $ / 1M tokens
$0.45
Output $ / 1M tokens
$2.25
Cached input $ / 1M
$0.070
Each benchmark shown separately with its own source

Benchmark results (12)

BenchmarkDomainScorevs best recordedSettingRunSource
CL-bench Lifelong-context13.2%
59%
——External ↗
CL-benchlong-context19.3%
69%
——External ↗
SimpleQA Verifiedknowledge34.3% ±1.5
45%
—27 Aug 2026Eval log ↗
Terminal Benchagents43.2% ±2.9
51%
agent: Terminus 227 Jan 2026External ↗
ARC-AGI-2reasoning11.8%
12%
——External ↗
Fiction.LiveBenchlong-context86.1%
89%
——External ↗
HLEknowledge24.4%
44%
——External ↗
WeirdMLcoding45.6% ±0.0
49%
——External ↗
SimpleBenchreasoning46.8%
57%
——External ↗
SWE-Bench verifiedcoding73.8% ±2.0
88%
—17 Feb 2026Eval log ↗
OSWorldagents63.3%
88%
agent: Kimi-K2.5—External ↗
ARC-AGIreasoning65.3%
66%
——External ↗

Source: Epoch AI Benchmarking Hub (CC BY 4.0). “External” rows are leaderboard results Epoch collects from third parties. Best reported setting per benchmark is shown.

API list price over time

Price history

$ per 1M tokens

9 recorded prices on 1 Feb 2026 (OpenRouter listing + Internet Archive snapshots); re-read on every data refresh, most recently 39m ago. Steps show when the price changed.

Events