Identity
- Status
- active
- Developer ID
moonshotai/Kimi-K2.6- Released
- Apr 20, 2026
- Version
- Kimi-K2.6
- License
- other
Moonshot AIKimi K2.6
Kimi-K2.6 is an open-weight vision-language model published by Moonshot AI.
- Token context
- 262K tokens
- Inputs
- Text, Image
- Outputs
- Text
- Released
- Apr 20, 2026
Providers & Pricing
Estimate workload cost →| Provider | Input / 1M | Output / 1M | Cache read / 1M | Cache write / 1M | Observed | Evidence |
|---|---|---|---|---|---|---|
 DeepinfraAvailable | $0.75 | $3.50 | $0.15 | — | Oct 19, 2026 | Provider-reported |
 OpenrouterAvailable | $0.65 | $3.41 | $0.15 | — | Oct 19, 2026 | Provider-reported |
Benchmarks
All benchmark results →| LiveBench2026-06-25-90f8d44da965 | 74.18 | 27,001 | 1,270 | kimi-k2.6-thinkingaverage per case | Recomputedwinner eligible | Third-party benchmark |
| LMArena Document Arenadocument-2026-09-13-a64ca3b12b6e | 1,450.74 | — | 11,291 | kimi-k2.6; 95% CI [1443.10467262, 1458.37474735]; votes 11291; rank 27unknown | Original sourcewinner eligible | Third-party benchmark |
| LMArena Text Arenatext-2026-10-15-a64ca3b12b6e | 1,455.42 | — | 40,013 | kimi-k2.6; 95% CI [1451.01237282, 1459.83173802]; votes 40013; rank 44unknown | Original sourcewinner eligible | Third-party benchmark |
| LMArena Vision Arenavision-2026-10-09-a64ca3b12b6e | 1,282.78 | — | 16,379 | kimi-k2.6; 95% CI [1275.98531996, 1289.57750584]; votes 16379; rank 33unknown | Original sourcewinner eligible | Third-party benchmark |
| ToneBench2026-10-17-10-task-e9c76358afb2 | 74.36 | 22,649 | 50 | Kimi K2.6 (high)average per case | Original sourcenot winner eligible | Third-party benchmark |
At a Glance
Model Facts
Capacity
- Token context
- 262K tokens
- Maximum output
- Unknown
- Total parameters
- 1.03T
- Active parameters
- 32B
- Knowledge cutoff
- Not reported
Interface and Access
- Inputs
- Text, Image
- Outputs
- Text
- API available
- Yes
- Open weights
- Yes
- Self-hostable
- Yes
- Reasoning
- Yes
- Vision
- Yes
- Tool calling
- Yes
- Capabilities
- chat, generation, reasoning, tools
This page represents one developer model product. Dated API snapshots, serving endpoints, pricing, regions, quantizations, and service tiers attach as versioned aliases or provider details and never create duplicate public model records.
Version Lineage
PredecessorUnknown
SuccessorsUnknown
Aliasesmoonshotai/Kimi-K2.6, moonshotai/Kimi-K2.6
Explicit Unknowns
- Maximum output
- Knowledge cutoff
Primary Evidence
Sources and Observation Date
moonshotai/Kimi-K2.6 developer model repositoryMoonshot AI · observed Aug 28, 2026moonshotai/Kimi-K2.6 README.md at 7eb5002f6aadMoonshot AI · observed Aug 28, 2026Deepinfra catalogDeepinfra · observed Oct 19, 2026OpenRouter moonshotai/kimi-k2.6 model APIOpenRouter · observed Aug 28, 2026moonshotai/Kimi-K2.6 Hugging Face inference-provider mappingHugging Face · observed Aug 28, 2026moonshotai/Kimi-K2.6 config.json at 7eb5002f6aadMoonshot AI · observed Aug 28, 2026moonshotai/Kimi-K2.6 immutable Hub API metadataMoonshot AI · observed Aug 28, 2026Kimi Research Blog release indexMoonshot AI · observed Aug 28, 2026moonshotai/Kimi-K2.6 tokenizer_config.json at 7eb5002f6aadMoonshot AI · observed Aug 28, 2026Fireworks Ai catalogFireworks Ai · observed Oct 19, 2026Hugging Face catalogHugging Face · observed Sep 2, 2026Openrouter catalogOpenrouter · observed Oct 19, 2026Together Ai catalogTogether Ai · observed Oct 19, 2026
Comparable Peers
Related Model Comparisons
| A | Pair | B | Context |
|---|---|---|---|
Kimi K2.6Moonshot AI | vs | Kimi K3Moonshot AI | family variantsembedding, image, text |
Kimi K2.6Moonshot AI | vs | cross-developer peerstext | |
Kimi K2.6Moonshot AI | vs |  Ministral 3 14B Reasoning 2512Mistral AI | cross-developer peersimage, text |
Kimi K2.6Moonshot AI | vs | Mistral Medium 3.5 128BMistral AI | cross-developer peersimage, text |
Kimi K2.6Moonshot AI | vs | Kimi K2 InstructMoonshot AI | family variantstext |
Kimi K2.6Moonshot AI | vs | Kimi K2.5Moonshot AI | family variantsembedding, image, text |
Kimi K2.6Moonshot AI | vs | Kimi K2.7 CodeMoonshot AI | family variantsimage, text |
Kimi K2.6Moonshot AI | vs | cross-developer peersimage, text | |
Kimi K2.6Moonshot AI | vs |  GLM 5.1Z.ai | cross-developer peerstext |
Questions
Kimi K2.6 FAQs
What is Kimi K2.6?+
Kimi-K2.6 is an open-weight vision-language model published by Moonshot AI. It is developed by Moonshot AI and its current sourced lifecycle status is active.
What inputs and outputs does Kimi K2.6 support?+
Kimi K2.6's current record lists Text and Image as input and Text as output.
How much context does Kimi K2.6 support?+
The current record lists a 262K-token context window and does not report a maximum output length.
Is Kimi K2.6 available through an API?+
Kimi K2.6 is marked as API-available. The directly sourced provider records currently include Deepinfra, Fireworks Ai, Hugging Face, Openrouter, and Together Ai.
How large is Kimi K2.6?+
The current record lists 1.03T total parameters and 32B active parameters. Its architecture is KimiK25ForConditionalGeneration.
What capabilities and tasks does Kimi K2.6 support?+
Kimi K2.6's recorded capabilities are chat, generation, reasoning, and tools. Its supported tasks are embedding, image, and text.
How much does Kimi K2.6 cost through an API?+
The lowest directly sourced prices currently attached to Kimi K2.6 are $0.65 per million input tokens through Openrouter and $3.41 per million output tokens through Openrouter. Prices are provider-specific and should be checked against each cited observation date.
Can Kimi K2.6 be self-hosted?+
Kimi K2.6 is an open-weight model. Self-hosting is supported, and the recorded license is other.
Does Kimi K2.6 have related versions or aliases?+
No predecessor is recorded. No successor is recorded. Recorded aliases are moonshotai/Kimi-K2.6.
Which models can Kimi K2.6 be compared with?+
The current compatible peer set includes Kimi K3, Qwen3.8 2.4T A95B, Ministral 3 14B Reasoning 2512, Mistral Medium 3.5 128B, and Kimi K2 Instruct. Compatibility requires shared sourced modalities and tasks; it does not imply that one model is better. Browse model comparisons →