Kimi K3.1 Pro / Highspeed
Overview
This page covers Kimi K3.1 Pro and Kimi K3.1 Pro Highspeed. Both variants offer the same intelligence and context window; their main differences are response speed and pricing.
Kimi K3.1 Pro is Moonshot AI's first K3.1-series flagship, a 3T-class model developed through further post-training of Kimi K3. Both Pro and its Highspeed variant combine native visual understanding and a 1M-token context window with improved coding, 3D creation, financial research, and multi-step agent capabilities.
Model Variants
| Model | API Model ID | Key Difference | Best For |
|---|---|---|---|
| Kimi K3.1 Pro | kimi-k3.1-pro | Lower input, output, and cache-write prices. | Cost-sensitive tasks. |
| Kimi K3.1 Pro Highspeed | kimi-k3.1-pro-highspeed | Approximately 2-3x the speed of the standard variant, according to Moonshot AI. | Interactive coding and latency-sensitive agent workflows. |
Key Features
- Software Engineering: Handles codebase fixes, debugging, and complete project builds, including interactive websites and games in tool-enabled environments.
- 3D and Spatial Reasoning: Interprets sketches, reference images, and text briefs to write Blender or CAD scripts for parametric models and scenes.
- Financial Research and Modeling: Supports cross-document analysis, numerical reasoning, and spreadsheet modeling through connected retrieval and execution tools.
- Adjustable Reasoning: Offers low, high, and max reasoning effort to balance depth, latency, and token usage.
- Long-Context Visual Understanding: Processes text, images, and video within a 1,048,576-token context window for substantial codebases and reference material.
Best Use Cases
- Repository Development: Investigate bugs, implement features, and iterate on larger projects using file, terminal, and testing tools.
- Interactive Web and Game Prototypes: Build frontend interfaces, Three.js scenes, and game experiences from written or visual references.
- Parametric 3D Design: Generate and refine scripts for mechanical parts, architectural models, furniture, and configurable products.
- Financial Analysis: Compare reports, trace supporting evidence, and prepare research or linked spreadsheet models with appropriate tools.
Capabilities and Limitations
The following capabilities and limitations apply to both model variants.
| Capability | Description |
|---|---|
| Reasoning | Always enabled, with low, high, and max effort; max is the default. |
| Input and Output | Text, image, and video input; text output, including code and structured JSON. |
| Context Window | 1,048,576 tokens, shared by input and generated output. |
| Max Output | 131,072 tokens by default; configurable up to 1,048,576, subject to remaining context capacity. |
| Tool Use | Function calling and dynamic tool loading for multi-step workflows. |
| Structured Output | JSON mode and JSON Schema output, including nested objects and arrays. |
| Streaming | Supports incremental reasoning and final-answer output. |
Known Limitations
- Sampling controls are fixed: temperature is 1.0 and top-p is 0.95; only one completion per request is supported.
- Multi-turn and tool-calling workflows must preserve complete assistant messages, including reasoning history.
- Producing 3D assets, running code, or editing spreadsheets requires suitable external tools and an execution environment.
Standard Pricing
The token prices below are shown in USD per 1 million tokens; web search is billed in USD per use.
| Model | Input (USD / 1M Tokens) | 5m Cache Write (USD / 1M Tokens) | 1h Cache Write (USD / 1M Tokens) | Cache Read (USD / 1M Tokens) | Output (USD / 1M Tokens) | Web Search (USD / use) |
|---|---|---|---|---|---|---|
| Kimi K3.1 Pro | $2.00 | $2.00 | $4.00 | $0.20 | $10.00 | - |
| Kimi K3.1 Pro Highspeed | $4.00 | $4.00 | $8.00 | $0.20 | $20.00 | - |
- Cache lifetimes are 5 minutes (5m, default) and 1 hour (1h). Cache hits renew the selected lifetime without another write charge.
- Input, cache-write, and cache-read rates apply to separate portions of input usage; the same tokens are not charged under multiple input categories.
Credits settlement: B.AI converts charges at 1 USD = 1,000,000 Credits and deducts Credits from the account balance.
Prices shown in the documentation are B.AI standard reference prices for base billing purposes. B.AI may provide lower actual usage costs through account benefits. Specific prices and account benefits are subject to the platform display and final billing records.