vs

One table to see how kimi-k3.1-pro (Moonshot AI) and hy3-preview (Tencent) differ. On EasyAPI both use the same endpoint and the same key; only the model name changes.

Quick verdict

  • hy3-preview has the lower input price, 92% cheaper than kimi-k3.1-pro.
  • hy3-preview has the lower output price, 95% cheaper than kimi-k3.1-pro.

Specs and pricing side by side

Dimensionkimi-k3.1-prohy3-preview
ProviderMoonshot AITencent
TypeTextText
Input price¥12.3871¥1.0205
Output price¥64.7943¥3.4017
Cache read¥1.2387¥0.3402
Cache write¥12.3871¥0
CapabilitiesTextText
Endpoints21

Platform prices per million tokens (image / video models use their own units); your invoice is authoritative.

kimi-k3.1-pro

Kimi K3.1 Pro is the first model in the Kimi K3.1 series, built for complex tasks such as long-horizon coding, 3D creation, and financial research. It excels at fixing issues in real codebases, multi-step tool use, and end-to-end project building, and can also handle game engine scripting, 3D modeling, and interactive scene generation. The model natively supports vision understanding and accepts video input, making it suitable for analyzing and generating mixed text-and-image content. Its thinking mode is always on, with reasoning effort adjustable on demand, making it useful for agent scenarios that require multi-turn planning, external tool collaboration, and result verification. It has a context window of about 1 million tokens and can be used for large codebases, long documents, and cross-document verification. The model supports function calling and structured output, and can be used in automated workflows with strict constraints on response format.

hy3-preview

Hy3 preview in high-reasoning mode, built for complex problem solving and multi-step agent workflows that need dependable execution.

FAQ

Which is cheaper, kimi-k3.1-pro or hy3-preview?
hy3-preview has the lower input price, 92% cheaper than kimi-k3.1-pro.
Can I call kimi-k3.1-pro and hy3-preview through the same API?
Yes. Both are served through EasyAPI's OpenAI-compatible API. Point base_url at EasyAPI and swap the model name; no separate vendor accounts needed.