ZMIME
Moonshot AI · Released Jan 27, 2026

Kimi K2.5

Earlier Kimi frontier model for long-context agents, coding, and multimodal work.

  • Open weights
  • Reasoning
  • Tool calling
  • Vision
  • Video input
Capability
148.8ECI · #73 of 158
Input
$0.60per 1M tokens
Output
$3.00per 1M tokens
Context
262K262K max output
Pricing

What it costs

Input
$0.60per million tokens
Output
$3.00per million tokens
Blended (3:1)
$1.20pricier than 62% of priced models

Typical monthly bills

Monthly usageEstimated cost
Side project2M input + 0.5M output tokens$2.70
Team assistant25M input + 5M output tokens$30.00
Production app250M input + 50M output tokens$300.00

Median of 36 paid API providers on models.dev.

Capability

Independent benchmarks

148.8Epoch Capabilities Index
#73 of 158 scored models
68Median 147174

The shaded band is Epoch AI’s confidence range (147.3–150.4); the tick marks the median scored model.

  • GPQA DiamondGraduate-level science questions87.6%
  • OTIS Mock AIME 2024–2025Competition mathematics92.2%
  • SWE-bench VerifiedFixing real GitHub issues73.8%
  • SimpleQA VerifiedShort factual questions34.3%

Scores from Epoch AI, run independently of Moonshot AI.

Specifications

The details

Lab
Moonshot AI
Released
Jan 27, 2026
Knowledge cutoff
Jan 2025
Context window
262,144 tokens
Max output
262,144 tokens
Inputs
Text, Images, Video
Output
Text
Reasoning
Always on
Tool calling
Yes
Structured output
No
Availability
36 API providerslisted on models.dev
Lab-reported

What Moonshot AI claims

Published by the lab at launch. Settings vary, so compare these only with care.

BenchmarkScoreSettingSource
SWE-Bench Verified70.8 resolved—Source
SWE-Atlas Codebase QnA13.1—Source
SWE-Atlas Refactoring20.95—Source
SWE-Atlas Test Writing25.77—Source
Same lab

More from Moonshot AI

Questions

About Kimi K2.5

How much does Kimi K2.5 cost?

Kimi K2.5 costs $0.60 input / $3.00 output per million tokens (median across 36 API providers). At a 3:1 input-to-output mix that is $1.20 per million tokens, more expensive than 62% of the 469 priced models we track.

What is the context window of Kimi K2.5?

Kimi K2.5 accepts up to 262,144 tokens per request and can write up to 262,144 tokens in one response.

How good is Kimi K2.5?

Epoch AI gives Kimi K2.5 a Capabilities Index score of 148.8 (likely range 147.3–150.4), ranking it #73 of 158 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 87.6%, OTIS Mock AIME 2024–2025 92.2%, SWE-bench Verified 73.8%, SimpleQA Verified 34.3%.

Is Kimi K2.5 open source?

Yes. Moonshot AI publishes the weights, so it can be downloaded and self-hosted.

What inputs does Kimi K2.5 support?

Kimi K2.5 accepts text, images and video and replies in text. It is a reasoning model, supports tool calling.

When was Kimi K2.5 released?

Moonshot AI released Kimi K2.5 on Jan 27, 2026. Its training data runs to Jan 2025.

What are the best alternatives to Kimi K2.5?

The closest current models from other labs on capability, price and release date are Qwen3.6 Plus (Alibaba (Qwen), ECI 148.9, $0.50 / $3.00), GPT-5.4 mini (OpenAI, ECI 149.4, $0.75 / $4.50), Grok 4.3 (xAI, ECI 149.9, $1.25 / $2.50) and GLM-5 (Z.ai (Zhipu), ECI 146.5, $1.00 / $3.20).