Moonshot AI (Kimi)

Kimi K2.5

Type
language model
Context
262K tokens
Max output
262K
Released
2 February 2026
Knowledge cutoff
January 2025
API string
moonshotai/Kimi-K2.5
Weights
open

Prices by provider

Provider Input, $ per 1M Output, $ per 1M In our data since
NanoGPT $0.3 $1.9 31 Jul 2026
crof $0.35 $1.7 31 Jul 2026
LLM Gateway $0.405 $1.98 31 Jul 2026
siliconflow-cn $0.45 $2.25 31 Jul 2026
meganova $0.45 $2.8 31 Jul 2026
DeepInfra $0.45 $2.25 31 Jul 2026
Kilo Code $0.45 $2.2 31 Jul 2026
SiliconFlow $0.45 $2.25 31 Jul 2026
auriko $0.5 $2.8 31 Jul 2026
tensorx $0.5 $2.8 2 Aug 2026
DigitalOcean Gradient AI $0.5 $2.7 31 Jul 2026
Nebius Token Factory $0.5 $2.5 31 Jul 2026
Together AI $0.5 $2.8 31 Jul 2026
neuralwatt $0.52 $2.59 31 Jul 2026
hyper $0.528 $2.79 1 Aug 2026
Cortecs $0.55 $2.76 31 Jul 2026
Venice AI $0.56 $3.5 31 Jul 2026
OpenRouter moonshotai/kimi-k2.5 $0.57 $2.85 31 Jul 2026
alibaba-cn kimi-k2.5 $0.574 $2.411 31 Jul 2026
ZenMux $0.58 $3.02 31 Jul 2026
opencode-go $0.6 $3 31 Jul 2026
daoxe $0.6 $3 31 Jul 2026
frogbot $0.6 $3 31 Jul 2026
moonshotai-cn $0.6 $3 31 Jul 2026
alibaba-cn kimi/kimi-k2.5 $0.6 $3 31 Jul 2026
novita-ai $0.6 $3 31 Jul 2026
hpc-ai $0.6 $3 31 Jul 2026
AIHubMix $0.6 $3 31 Jul 2026
Abacus.AI $0.6 $3 31 Jul 2026
Amazon Bedrock bedrock/moonshotai.kimi-k2.5 $0.6 $3.03 31 Jul 2026
Amazon Bedrock bedrock/us-east-1/moonshotai.kimi-k2.5 $0.6 $3 31 Jul 2026
Azure Cognitive Services $0.6 $3 31 Jul 2026
Baseten $0.6 $3 31 Jul 2026
Cloudflare AI Gateway $0.6 $3 31 Jul 2026
Fireworks AI $0.6 $3 1 Aug 2026
Hugging Face $0.6 $3 31 Jul 2026
Jiekou.AI $0.6 $3 31 Jul 2026
Merge Gateway $0.6 $3 31 Jul 2026
Microsoft Foundry $0.6 $3 31 Jul 2026
Moonshot AI $0.6 $3 31 Jul 2026
OpenCode Zen $0.6 $3 31 Jul 2026
OpenRouter openrouter/moonshotai/kimi-k2.5 $0.6 $3 31 Jul 2026
OrcaRouter $0.6 $3 31 Jul 2026
Poe $0.6 $3 31 Jul 2026
Vercel $0.6 $3 31 Jul 2026
Weights & Biases $0.6 $3 31 Jul 2026
CrossModel $0.62 $3.3 31 Jul 2026
Amazon Bedrock bedrock/ap-northeast-1/moonshotai.kimi-k2.5 $0.72 $3.6 31 Jul 2026

Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.

A single provider appears several times if it sells this model under different identifiers and at different prices — for example, at different weight precision. The identifier is shown next to the name. Identical offers that arrived from two sources under different spellings are merged into one row.

The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.

Measurement results

Task set Result Run conditions Measured by
Mock AIME 2024–2025 — olympiad problems 92.19 % · with a tuned harness Epoch evaluations
Fiction.LiveBench — holding a long context 86.1 % Fiction.live leaderboard
GPQA Diamond — graduate-level questions 83.47 % · with a tuned harness Epoch evaluations
SWE-bench Verified — fixing bugs in repositories 73.76 % · with a tuned harness Epoch evaluations
ARC-AGI — generalising to unseen patterns 65.33 % · with a tuned harness
OSWorld — working inside an operating system, first version 63.3 % · with a tuned harness OS World Website
WeirdML — unusual machine learning tasks 45.6 % WeirdML Leaderboard
Terminal-Bench — working in the command line 43.2 % · with a tuned harness
SimpleBench — trick questions 36.16 % SimpleBench Leaderboard
SimpleQA Verified — factual accuracy 33.9 % Epoch evaluations
CursorBench — edits in the editor 31.9 % https://cursor.com/cursorbench
Humanity’s Last Exam — expert-level questions 20.56 %
APEX-Agents 14.4 %
ARC-AGI-2 11.81 %
Chess puzzles 7.41 % Epoch evaluations