moonshotai
MoonshotAI: Kimi K2.5
moonshotai/kimi-k2.5
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
MoonshotAI: Kimi K2.5 (moonshotai/kimi-k2.5) is an open-source, natively multimodal mixture-of-experts model described by Moonshot AI as accepting image and text inputs, producing text, and offering instant and thinking modes plus coordinated agent-swarm execution. Its model card reports results across visual, video, document, math, coding, and agentic evaluations, while OpenRouter lists a 256K-token context and tool or structured-output controls. Because the evidence is mainly vendor documentation plus a catalog entry, validate performance, cost, reliability, and safety on intended workloads.[1][2]
Strengths
- The vendor describes native multimodality with a MoonViT vision encoder, visual knowledge, cross-modal reasoning, and image-grounded tool use.[2]
- Moonshot AI reports instant and thinking modes and a self-directed swarm scheme that divides complex tasks into parallel, domain-specific subtasks.[2]
- The OpenRouter catalog lists a 256K-token context, image and text input, text output, tools, structured outputs, and reasoning controls.[1]
Limitations
- Evidence is limited to official model documentation and an OpenRouter catalog entry; no independent card, reproduction, or workload-specific testing was supplied.[1][2]
- The catalog lists image and text inputs but only text outputs; although the vendor evaluates video, direct video input support is not specified.[1][2]
- No hardware requirements, deployment limits, knowledge cutoff, or safety and bias evaluations are stated; review the modified-MIT license before reuse.[1][2]
Where to use it
- Evaluate for UI-to-code conversion, visual debugging, and code generation from screenshots or visual specifications.[1][2]
- Test in agentic workflows needing tool calls and parallel decomposition, while validating orchestration behavior in the target stack.[1][2]
- Consider long-context multimodal assistants over documents, images, and visually grounded questions, subject to the text-only output constraint.[1][2]
Sources
- [1] OpenRouter catalog · MoonshotAI: Kimi K2.5Sources checked: Nov 13, 2026
- [2] Model documentation · MoonshotAI: Kimi K2.5Sources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm. Built on Kimi K2 with continued pretraining over approximately 15T mixed...
Specifications
- Model ID
- moonshotai/kimi-k2.5
- Canonical ID
- moonshotai/kimi-k2.5-0127
- Creator
- moonshotai
- Added to OpenRouter
- 27.01.2026
- Context length
- 262,144
- Modality
- text+image->text
- Input modalities
- text, image
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.500 / out $2.5
- 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 | BB — next 30% of the cohort | — |
| Agents | No measurement | 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 | 46.8 | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | agents · androidnativewin rate 57.8% · rank 33 | 1,105Elo | CC — below the B band | 13.11.2026 |
| Design Arena | agents · fullstackwin rate 54.4% · rank 30 | 1,091Elo | CC — below the B band | 13.11.2026 |
| Design Arena | agents · godotgamedevwin rate 59.8% · rank 9 | 1,219Elo | AA — next 20% of the cohort | 13.11.2026 |
| Design Arena | agents · mobileappswin rate 54.7% · rank 33 | 1,142Elo | CC — below the B band | 13.11.2026 |
| Design Arena | agents · webappswin rate 51.6% · rank 34 | 1,117Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · 3dwin rate 53.1% · rank 55 | 1,188Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · asciiartwin rate 46.5% · rank 37 | 1,160Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · codecategorieswin rate 54.1% · rank 52 | 1,227Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · datavizwin rate 51.3% · rank 53 | 1,217Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · gamedevwin rate 53.4% · rank 58 | 1,197Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · svgwin rate 48.4% · rank 54 | 1,141Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · uicomponentwin rate 53.6% · rank 50 | 1,229Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · websitewin rate 55.2% · rank 49 | 1,243Elo | BB — next 30% 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) |
|---|---|---|---|---|---|---|
| AtlasCloudint4 | 262,144 | $0.490 | $2.5 | — | — | — |
| SiliconFlowint4 | 262,144 | $0.500 | $2.5 | — | — | 100% |
| Venice | 256,000 | $0.532 | $3.32 | — | — | 100% |
| Novita | 262,144 | $0.570 | $2.85 | — | — | 100% |
| Amazon Bedrock | 262,144 | $0.600 | $3 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.