Model

Kimi K2.5

Name: Kimi K2.5
Brand: moonshotai

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

45 providers · 50 listingsreleased Jan 1, 2026knowledge cutoff 2025-01open weights

$0.3

lowest input /M · NanoGPT

$1.90

lowest output /M · NanoGPT

262K

context window

262K

max output

Compare providers (sorted by listed input price)

37 of 45 providers publish a price for Kimi K2.5. The cheapest full-context option is NanoGPT at $0.3/$1.9 per M. Listed prices span 2.0× for identical weights.

			Cache read /M	Cache write /M		Capabilities	Status
NanoGPT moonshotai/kimi-k2.5	$0.3	$1.90	$0.15	—	256K	no reasoningtoolsno structuredvision		Jan 1, 2026
LLM Gatewayvariant embercloud/kimi-k2.5	$0.405	$1.98	$0.225	—	262K	no reasoningtoolsno structuredno vision		Jan 1, 2026
DevPass (LLM Gateway) kimi-k2.5	$0.405	$1.98	$0.225	—	262K	reasoningtoolsstructuredvision		Jan 1, 2026
SiliconFlow (China) Pro/moonshotai/Kimi-K2.5	$0.45	$2.25	$0.07	—	262K	reasoningtoolsstructuredno vision		Jan 27, 2026
OpenRouter moonshotai/kimi-k2.5	$0.45	$2.25	$0.07	—	262K	reasoningtoolsstructuredvision		Jan 1, 2026
SiliconFlow moonshotai/Kimi-K2.5	$0.45	$2.25	$0.07	—	262K	reasoningtoolsstructuredno vision		Jan 27, 2026
Meganova moonshotai/Kimi-K2.5	$0.45	$2.80	—	—	262K	reasoningtoolsno structuredno vision		Jan 27, 2026
Eden AIvariant deepinfra/moonshotai/Kimi-K2.5	$0.45	$2.25	$0.07	—	262K	reasoningtoolsstructuredvision		Jan 1, 2026
Deep Infra moonshotai/Kimi-K2.5	$0.45	$2.25	$0.07	—	262K	reasoningtoolsstructuredvision	deprecated	Jan 27, 2026
Cortecs kimi-k2.5	$0.495	$2.77	$0.124	—	262K	reasoningtoolsstructuredvision		Jan 1, 2026
DigitalOcean kimi-k2.5	$0.5	$2.70	$0.203	—	262K	reasoningtoolsstructuredvision		Apr 16, 2026
Auriko kimi-k2.5	$0.5	$2.80	—	—	262K	reasoningtoolsstructuredvision		Jan 1, 2026
TensorX moonshotai/kimi-k2.5	$0.5	$2.80	$0.125	$0.625	262K	reasoningtoolsstructuredvision		Jan 1, 2026
Eden AIvariant tensorx/moonshotai/kimi-k2.5	$0.5	$2.80	$0.125	—	262K	reasoningtoolsno structuredvision		Jan 1, 2026
Together AI moonshotai/Kimi-K2.5	$0.5	$2.80	—	—	262K	reasoningtoolsno structuredno vision	deprecated	Jan 27, 2026
Charm Hyper kimi-k2.5	$0.558	$2.94	$0.279	—	262K	reasoningtoolsstructuredno vision		Jul 22, 2026
Venice AI kimi-k2-5	$0.56	$3.50	$0.22	—	256K	reasoningtoolsstructuredvision		Jun 11, 2026
LLM Gatewayvariant alibaba/kimi-k2.5	$0.574	$3.01	—	—	262K	reasoningtoolsno structuredvision		Jan 1, 2026
Alibaba (China)variant kimi-k2.5	$0.574	$2.41	—	—	262K	reasoningtoolsno structuredno vision		Jan 27, 2026
ZenMux moonshotai/kimi-k2.5	$0.58	$3.02	$0.1	—	262K	reasoningtoolsno structuredvision		Jan 27, 2026

Best price history

Daily lowest published price across live providers. A lower line is better for the buyer.

Best input /M: $0.3▼ 14.3%
Best output /M: $1.90▲ 11.8%
Provider coverage: 50; 57 at the start · 24 snapshots

Best input priceBest output priceAug 21, 2026 → Sep 14, 2026

Price badge

![Kimi K2.5 price](https://www.vaanalytics.in/badge/moonshotai/kimi-k2.5.svg)

https://www.vaanalytics.in/badge/moonshotai/kimi-k2.5.svg

Live SVG, regenerated on every hourly sync — paste it into a README to always show the current cheapest listed price. Updates when prices change; no build step needed on your side.

Recent changes

Changelog →

contextvia amazon-bedrock · Sep 13, 2026
- description: Kimi multimodal agent model for visual understanding, coding, and planning → Earlier Kimi frontier model for long-context agents, coding, and multimodal work
- family: kimi → kimi-k2
- attachment: false → true
- knowledge: — → 2025-01
- +2 more
repricedvia hyper · Sep 11, 2026
- cache_read: — → $0.279
- cache_write: $0.279 → —
new modelvia 302ai · Sep 9, 2026
deprecatedvia deepinfra · Sep 8, 2026
- status: — → deprecated
repricedvia hyper · Sep 8, 2026
- input: $0.514 → $0.558
- output: $2.75 → $2.94
- cache_write: $0.257 → $0.279
removedvia moonshotai · Sep 7, 2026

Benchmarks

SWE-Bench Verified ↗(resolved)report ↗Independent70.8
SWE-Atlas Codebase QnA(score)report ↗Independent13.1
SWE-Atlas Refactoring(score)report ↗Independent20.95
SWE-Atlas Test Writing(score)report ↗Independent25.77

Scores a lab published about its own model are claims, not measurements. Rankings elsewhere on this site use independently measured scores only.

Weights

Licence unclearother1.03T params

No machine-readable licence on the model card. Check it before you rely on this. Weights download without an access request.

licence and gating from the Hugging Face model card ↗ — the card is authoritative, this is a summary

Hugging Face

Prices are per million tokens (USD). “—” means the provider does not publicly list a price for this model. Data from models.dev; verify with the provider before purchasing.