Kimi k1 / k-series is a Long-context flagship from Moonshot AI (China).
One-line fit: best used for Huge docs, research assistants.
Key specifications
| Provider | Moonshot AI |
|---|---|
| Model | Kimi k1 / k-series |
| Type | Long-context flagship |
| Context window | Up to millions-class claims by generation |
| Pricing (indicative) | App + API |
| Company category | Frontier Lab |
| Region | China |
| Official site | https://www.moonshot.cn |
Benchmarks and public standing
Long-context needle and agent demos
Treat scores as directional. SWE-bench, Terminal-Bench, LMSYS Arena, Artificial Analysis, and vendor cards use different harnesses and are not always comparable 1:1. Re-check the latest official model card before decisions.
What this model is good at
Huge docs, research assistants
This aligns with Moonshot AI’s broader strengths:
- Ultra-long context
- Browser/agent products
- CN consumer UX
Limitations and watch-outs
Global API ecosystem smaller than Qwen
- Rate limits, regional availability, and data retention policies vary by plan.
- Agent harness quality (Cursor, Claude Code, Codex, custom tools) can change outcomes more than raw model Elo.
- Open-weight availability (if any) is separate from hosted API quality and safety filters.
Ideal users
- Product engineers shipping features that match: Huge docs, research assistants
- Teams standardizing on the Moonshot AI ecosystem
- Agent builders who need a Long-context flagship
When to pick something else
- Cheaper volume: compare lower tiers from the same lab or open Chinese/EU alternatives.
- Maximum hard SWE thoughtfulness: compare Claude Fable-class and other coding flagships.
- Giant multimodal corpora / Workspace: compare Gemini-class models.
- Self-host / open weights: Llama, Qwen, DeepSeek, GLM, Mistral open lines.
Parent company
Moonshot AI – Kimi assistant famous for very long context and agentic browsing products.
Related models from Moonshot AI
Data is curated for AIForumSphere model directory mid-2026 directional figures.
