Project AURIONAI Use-Cases Research & Innovation
multimodal LLM

Kimi K2.5

multimodal LLM from Moonshot AI.

Moonshot AI
Levelmodel
Classmultimodal LLM
Parameters1T
Contextnot in the notes
ReleasedJan 2026
Opennessopen weights restricted
Runs onnot in the notes
Statusactive

Benchmark Results

vendor-reported
vendor-reported
76.8
050100

Score in %. Higher is better.

Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

Lineage

successorfinetune
Composer 2 to Composer 2.5: successor, medium confidenceKimi K2-0905 to Kimi K2 Thinking: successor, medium confidenceKimi K2.5 to Composer 2.5: finetune, medium confidenceKimi K2.5 to Kimi K2.6: successor, medium confidenceKimi K2.6 to Kimi K2.7 Code: finetune, high confidenceKimi K2.6 to Kimi K3: successor, medium confidenceKimi K2 Thinking to Kimi K2.5: successor, medium confidenceKimi K2-0905Kimi K2-0905Moonshot AIKimi K2 ThinkingKimi K2 ThinkingMoonshot AIKimi K2.5Kimi K2.5Moonshot AI, Jan 2026Composer 2Composer 2Anysphere (Cursor), Mar 2026Kimi K2.6Kimi K2.6Moonshot AI, Apr 2026Composer 2.5Composer 2.5Anysphere (Cursor), May 2026Kimi K2.7 CodeKimi K2.7 CodeMoonshot AI, Jun 2026Kimi K3Kimi K3Moonshot AI, Jul 2026
Anysphere (Cursor)Composer 2.5 is a child (finetune)medium confidencepress claim
Moonshot AIKimi K2.6 is a child (successor)medium confidencenotes
Moonshot AIKimi K2 Thinking is the parent (successor)medium confidencenotes
Where this comes from
Parent and child relations from model cards, Hugging Face base_model fields and the genealogy tables of the research notes. The tree shows two generations each way, 8 models. Confidence and evidence are the notes' own.

R&D Topics

Identity

Developer
Moonshot AI
Family
Kimi
Released
Jan 2026 sourcessecondary report
Weights published
30 Apr 2026
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
MoE
Total parameters
1T sources
Active parameters
–
Checkpoint parameters
–
Note
HF dates listed as May 19 / Apr 30 for K2.6 / K2.5 (year not stated; HF update dates)
Context length
not in the notes
Input
text
Output
text
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
open weights restricted
Licence
Modified MIT (Moonshot AI) sources
Access routes
not in the notes
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
not in the notes
Footprint
not in the notes
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
not in the notes
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Industry Use Cases

No industry use case links this model.

Sources

  1. Kimi (AI) - Wikipedia en.wikipedia.org https://en.wikipedia.org/wiki/Kimi_(chatbot)
  2. huggingface.co/moonshotai https://huggingface.co/moonshotai
  3. huggingface.co/moonshotai https://huggingface.co/moonshotai
  4. leaderboard.steel.dev/leaderboards/swe-bench-verified.md https://leaderboard.steel.dev/leaderboards/swe-bench-verified.md
  5. Kimi K3 Open Weights Ship: 2.8T Parameters, 1.4 TB to Run rits.shanghai.nyu.edu https://rits.shanghai.nyu.edu/ai/kimi-k3-open-weights-ship-2-8t-parameters-1-4-tb-to-run/

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Kimi K2.5", https://aurion.semai.id/models/moonshot-ai-kimi-k2-5/, records of October 2026.