moonshotai
MoonshotAI: Kimi K2.6
moonshotai/kimi-k2.6
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Kimi K2.6 is Moonshot AI's open-source, natively multimodal agentic model, accepting text and image input and producing text. The vendor's model card describes a 1T-parameter mixture-of-experts design with 32B activated parameters, a MoonViT vision encoder, and a roughly 256K-token context, matching the 262,144 tokens listed in the OpenRouter catalog. Moonshot AI positions it for long-horizon coding, code-driven interface generation, and multi-agent orchestration. Its evaluation table is vendor-reported, and the catalog lists only limited third-party index measurements, so independent workload validation is advised.[1][2]
Strengths
- Natively multimodal: accepts text and image inputs and returns text, with a context window of about 256K tokens.[1][2]
- The vendor reports strong capabilities in long-horizon, end-to-end coding across Python, Rust, and Go, plus code-driven UI and full-stack generation.[2]
- The catalog lists tool calling, parallel tool calls, structured outputs, and an optionally enabled reasoning mode.[1]
Limitations
- Most capability and benchmark evidence is vendor-reported; the catalog's third-party index data is sparse and does not independently confirm the agentic claims.[1][2]
- No knowledge cutoff is stated, and the license is a modified MIT variant that should be reviewed before deployment.[1][2]
- Evidence is limited to the catalog entry and the vendor card; teams should validate quality, cost, and latency on their own workloads.[1][2]
Where to use it
- Worth evaluating for long-running, end-to-end coding and terminal tasks that span multiple languages.[2]
- A candidate for prototypes that turn prompts or visual inputs into interfaces or lightweight full-stack workflows.[2]
- Potentially suited to orchestrated multi-agent pipelines producing documents, slides, spreadsheets, or websites.[1][2]
Sources
- [1] OpenRouter catalog · MoonshotAI: Kimi K2.6Sources checked: Nov 13, 2026
- [2] Model documentation · MoonshotAI: Kimi K2.6Sources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Kimi K2.6 is Moonshot AI's next-generation multimodal model, designed for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration. It handles complex end-to-end coding tasks across Python, Rust, and Go, and...
Specifications
- Model ID
- moonshotai/kimi-k2.6
- Canonical ID
- moonshotai/kimi-k2.6-20260420
- Creator
- moonshotai
- Added to OpenRouter
- 20.04.2026
- Context length
- 262,144
- Modality
- text+image->text
- Input modalities
- text, image
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.650 / out $3.41
- Knowledge cutoff
- —
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Coding | AA — next 20% of the cohort | — |
| Agents | BB — next 30% of the cohort | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Document analysis | No measurement | — |
| Vision | No measurement | input: image |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Coding index | 61.8 | AA — next 20% of the cohort | 13.11.2026 |
| Artificial Analysis | Agentic index | 20.5 | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | agents · agenticgamedevwin rate 47.5% · rank 21 | 1,088Elo | CC — below the B band | 13.11.2026 |
| Design Arena | agents · agentichtmlslideswin rate 59% · rank 2 | 1,248Elo | —Not rated | 13.11.2026 |
| Design Arena | agents · agenticslideswin rate 45.8% · rank 4 | 1,187Elo | —Not rated | 13.11.2026 |
| Design Arena | agents · agenticslides(html)win rate 59.2% · rank 2 | 1,252Elo | —Not rated | 13.11.2026 |
| Design Arena | agents · agenticslides(python-pptx)win rate 45.5% · rank 4 | 1,186Elo | —Not rated | 13.11.2026 |
| Design Arena | agents · androidnativewin rate 49.2% · rank 27 | 1,151Elo | CC — below the B band | 13.11.2026 |
| Design Arena | agents · fullstackwin rate 49.4% · rank 27 | 1,134Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | agents · godotgamedevwin rate 46.8% · rank 28 | 1,103Elo | CC — below the B band | 13.11.2026 |
| Design Arena | agents · htmlslideswin rate 52.3% · rank 13 | 1,081Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | agents · mobileappswin rate 47.6% · rank 32 | 1,149Elo | CC — below the B band | 13.11.2026 |
| Design Arena | agents · pptxslideswin rate 44.3% · rank 4 | 1,181Elo | —Not rated | 13.11.2026 |
| Design Arena | agents · python-pptxslideswin rate 42.1% · rank 14 | 1,180Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | agents · spreadsheetwin rate 54.8% · rank 8 | 1,276Elo | AA — next 20% of the cohort | 13.11.2026 |
| Design Arena | agents · webappswin rate 59.3% · rank 11 | 1,268Elo | AA — next 20% of the cohort | 13.11.2026 |
| Design Arena | models · 3dwin rate 57.6% · rank 31 | 1,250Elo | AA — next 20% of the cohort | 13.11.2026 |
| Design Arena | models · asciiartwin rate 46.1% · rank 41 | 1,152Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · codecategorieswin rate 55% · rank 34 | 1,263Elo | AA — next 20% of the cohort | 13.11.2026 |
| Design Arena | models · datavizwin rate 52.1% · rank 37 | 1,247Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · gamedevwin rate 54.8% · rank 42 | 1,240Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · svgwin rate 51.2% · rank 40 | 1,172Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · uicomponentwin rate 55.8% · rank 39 | 1,258Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · websitewin rate 54.6% · rank 32 | 1,264Elo | AA — next 20% of the cohort | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| Baidufp4 | 262,144 | $0.390 | $1.64 | — | — | 100% |
| Chutesint4 | 262,144 | $0.390 | $1.8 | — | — | 100% |
| DigitalOcean | 262,144 | $0.570 | $2.4 | — | — | 100% |
| StreamLakefp8 | 256,000 | $0.598 | $2.52 | — | — | — |
| CoreWeavefp4 | 262,144 | $0.650 | $3.41 | — | — | 100% |
| Crusoebf16 | 262,144 | $0.700 | $3.5 | — | — | 100% |
| DeepInfrafp4 | 262,144 | $0.750 | $3.5 | — | — | 100% |
| Veniceint4 | 256,000 | $0.750 | $3.5 | — | — | — |
| Parasailint4 | 262,144 | $0.750 | $3.5 | — | — | 100% |
| SiliconFlowfp8 | 262,144 | $0.770 | $3.4 | — | — | — |
| Inceptronint4 | 262,144 | $0.798 | $3.75 | — | — | 100% |
| Novita | 262,144 | $0.800 | $3.4 | — | — | 100% |
| AtlasCloudint4 | 262,144 | $0.950 | $4 | — | — | — |
| Cloudflare | 262,144 | $0.950 | $4 | — | — | — |
| Moonshot AI | 262,144 | $0.950 | $4 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.