moonshotai
MoonshotAI: Kimi K2 0905
moonshotai/kimi-k2-0905
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Moonshot AI's Kimi K2 0905 is the September update of Kimi K2, described in its model documentation as a mixture-of-experts language model with one trillion total parameters and 32 billion activated per token. The vendor highlights stronger agentic and frontend coding, tool calling, and a context window enlarged to 256k tokens. Vendor-run benchmarks report gains over the previous release on software-engineering suites, but these figures are self-reported. OpenRouter's catalog lists text-only input and output with structured-output and tool support. Independent third-party evaluations were not included in this evidence pack.[1][2]
Strengths
- Vendor documentation describes a one-trillion-parameter MoE design with 32B activated parameters per token and a 256k-token context window, doubled from the prior release.[1][2]
- The vendor reports improved agentic and frontend coding plus strong tool calling; OpenRouter's catalog lists tools, tool_choice, structured outputs and seed support.[1][2]
- Vendor-run coding benchmarks report higher accuracy than the earlier K2 version, but these are self-reported results, not independent measurements.[2]
Limitations
- Evidence comes only from the vendor's model card and the OpenRouter catalog; no independent benchmark or review was supplied, so claims should be validated on your own workloads.[1][2]
- OpenRouter lists text as the only input and output modality; other modalities are not documented.[1]
- OpenRouter reports a knowledge cutoff at the end of 2024 and no dedicated reasoning mode; the license is a modified MIT variant worth reviewing before use.[1][2]
Where to use it
- Agentic coding or software-maintenance assistants, where the vendor reports its strongest results; a suggested fit pending tests on your own repositories.[2]
- Long-document analysis or multi-step tasks that could benefit from the documented 256k-token context window.[1][2]
- Tool-using chat applications via OpenAI- or Anthropic-compatible APIs or inference engines such as vLLM and SGLang.[2]
Sources
- [1] OpenRouter catalog · MoonshotAI: Kimi K2 0905Sources checked: Nov 13, 2026
- [2] Model documentation · MoonshotAI: Kimi K2 0905Sources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Kimi K2 0905 is the September update of [Kimi K2 0711](moonshotai/kimi-k2). It is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32...
Specifications
- Model ID
- moonshotai/kimi-k2-0905
- Creator
- moonshotai
- Added to OpenRouter
- 04.09.2025
- Context length
- 262,144
- Modality
- text->text
- Input modalities
- text
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.600 / out $2.5
- Knowledge cutoff
- 31.12.2024
- 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 | — |
| Coding | No measurement | — |
| Agents | No measurement | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | — |
| Document analysis | No measurement | — |
| Vision | No measurement | — |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Design Arena | models · codecategorieswin rate 48.5% · rank 112 | 1,088Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · websitewin rate 48.3% · rank 114 | 1,100Elo | CC — below the B band | 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) |
|---|---|---|---|---|---|---|
| Novitafp8 | 262,144 | $0.600 | $2.5 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.