# Moonshot AI vs StreamLake

> Two Chinese coding contenders. Kimi K3 is open-weight with independent benchmark results; StreamLake's KAT-Coder is proprietary and sold on a subscription.

Canonical: https://www.subconscious.dev/compare/moonshot-ai-vs-streamlake · By The Subconscious Team · Updated September 30, 2026

## How they compare

Moonshot and StreamLake both pitch agentic coding, with different models and terms. Kimi K3 is open-weight, with a 1M window and independent results that include 93.4% on SWE-bench Verified from Vals AI and third place on the Artificial Analysis Intelligence Index. StreamLake's KAT-Coder-Pro V2.5 is proprietary, trained by Kuaishou's KwaiKAT team with large-scale agentic reinforcement learning for repository-level work, and its listing relies on StreamLake's own description rather than independent scores. K3 costs $3 in and $15 out. KAT-Coder comes per token or on the KwaiKAT Coding Plan.

The tooling paths differ. StreamLake offers OpenAI-protocol endpoints and a Claude-protocol proxy that drops into Claude Code or OpenClaw. Moonshot offers an OpenAI-compatible endpoint, Kimi Code in the terminal, OpenRouter and Cloudflare Workers AI. StreamLake's pricing and docs lead with China and yuan, and its data residency in China rules it out for many US and EU buyers. Moonshot's hurdles are K3's speed, around 33 tokens per second, and the capacity limits seen at launch. Open weights also give K3 users an exit path that KAT-Coder does not.

## What each one does

### Moonshot AI

Moonshot AI is the Beijing lab behind the Kimi models. Its flagship Kimi K3 launched July 16, 2026 as a 2.8 trillion parameter mixture-of-experts model that activates 16 of 896 experts per token, with native vision and a 1M token context. It is the first open model in the 3T class, and full weights landed on Hugging Face on July 27. The hosted API costs $3 in and $15 out per million tokens, with cached input at $0.30, and it runs through an OpenAI-compatible endpoint, Kimi Code in the terminal, OpenRouter and Cloudflare Workers AI.

### StreamLake

StreamLake is the AI cloud brand of Kuaishou, the Chinese short-video company behind the Kling video models. It sells model-as-a-service inference and bare-metal compute to internet businesses, drawing on the infrastructure Kuaishou built to serve video at massive scale. Its developer site offers APIs, SDKs and integration guides aimed at taking teams from testing to production.

## Which is best, and when

### Choose Moonshot AI for

- Teams that want open weights they can self-host later
- Benchmark-backed coding quality on hard tasks
- Distribution through OpenRouter and Cloudflare Workers AI

### Choose StreamLake for

- Subscription pricing for heavy KAT-Coder use
- Running KAT-Coder inside Claude Code through its proxy
- Buyers in China who also want bare-metal capacity

## At a glance

| Attribute | Moonshot AI | StreamLake |
|---|---|---|
| Model access | Open weights, custom license | Proprietary coding models |
| Flagship models | Kimi K3, Kimi K2.6 | KAT-Coder-Pro V2.5, KAT-Coder-Air |
| Speed | ~33 tok/s on Kimi K3 | - |
| Price | $3 in, $15 out (Kimi K3) | Per token or KwaiKAT Coding Plan |
| Customization | Open weights to fine-tune | - |
| Deployment | API, Kimi Code, OpenRouter | MaaS API, bare metal |
| Long context | 1M | - |

## FAQ

### What is the difference between Moonshot AI and StreamLake?

Two Chinese coding contenders. Kimi K3 is open-weight with independent benchmark results; StreamLake's KAT-Coder is proprietary and sold on a subscription.

### When should I choose Moonshot AI over StreamLake?

Teams that want open weights they can self-host later; Benchmark-backed coding quality on hard tasks; Distribution through OpenRouter and Cloudflare Workers AI.

### When should I choose StreamLake over Moonshot AI?

Subscription pricing for heavy KAT-Coder use; Running KAT-Coder inside Claude Code through its proxy; Buyers in China who also want bare-metal capacity.

### Is Moonshot AI or StreamLake cheaper?

Moonshot AI: $3 in, $15 out (Kimi K3). StreamLake: Per token or KwaiKAT Coding Plan. The cheaper choice depends on the model and workload.

## Running long-horizon agents?

If your agents run past 200K tokens, compare both against Subconscious: [Subconscious vs Moonshot AI](https://www.subconscious.dev/compare/subconscious-vs-moonshot-ai.md), [Subconscious vs StreamLake](https://www.subconscious.dev/compare/subconscious-vs-streamlake.md).

Full profiles: [Moonshot AI](https://www.subconscious.dev/providers/moonshot-ai.md), [StreamLake](https://www.subconscious.dev/providers/streamlake.md).