MDL
Kimi K2.5
Kimi
Verified against Artificial Analysis · Jul 13, 2026
Capability
89
percentile index
Price /M
$0.6 / $3
in / out
Context
262.1K
tokens
Avg score
61.8
7 benchmarks
AA Intelligence
38.1
index
Speed
49 tok/s
TTFT 1.1s
Released
Jan 27, 2026
About
Kimi K2.5 is a open-weights model in the Kimi family from Kimi. Benchmarked on 7 evals, averaging 61.8.
Benchmark scores · 7
avg 61.8 — higher is betterThe research behind this model
all papers mentioning it →SkillAdaptor: Self-Adapting Skills for LLM Agents from Trajectories▲ 35 on HF · May 31, 2026HRBench: Benchmarking and Understanding Thinking-Mode Switch Strategies in Hybrid-Reasoning LLMs▲ 15 on HF · May 27, 2026ProCUA-SFT Technical Report▲ 9 on HF · Jun 15, 2026PetroBench: A Benchmark for Large Language Models in Petroleum EngineeringMay 27, 2026ETCHR: Editing To Clarify and Harness ReasoningMay 22, 2026
Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.
Performance & price
Output speed49 tok/s
Latency (TTFT)1.1s
First answer1m 2s
Price in / out$0.6 / $3 per 1M
Blended price$1.2 / 1M
Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →
History
full ledger →Capability over time
Append-only ledger — every observed change to this model's numbers.
API providers
Artificial Analysis$3/M