Approved models
10
Records published from this collection
Provider information
lab
Review approach
Verify each model’s source, terms, capabilities, and current availability
Moonshot AI models in this directory cover any-to-any, text-to-text ai models. Start with the task you need to solve, then open a model profile to check its recorded inputs, outputs, licensing and deployment options.
Use inputs that resemble deployment traffic, including difficult and out-of-domain cases. Compare output correctness, the response format, latency and memory requirements with a simple baseline under the same conditions. Open-weight checkpoints need a runtime and enough memory for the exact format you select; a repository listing alone does not provide a hosted API.
Check whether two entries describe distinct models or different hosted routes for the same model. Keep the checkpoint, host, test settings and evaluation date together. Missing prices and limits mean the directory has no recorded figure, not that use is free or unlimited.
Browse the AI model directory or compare two models.
Review each record individually; collection membership does not imply a recommendation or guarantee of availability.
| Kimi K3 Moonshot AI | 80.8 | 1.05M | $3.00 | $15.00 | — |
| Kimi K2.8 Preview Moonshot AI | — | 1.05M | — | — | — |
| Kimi K2.7 Code Moonshot AI | 74.9 | 262.14K | $0.95 | $4.00 | — |
| Kimi K2.7 Code Highspeed Moonshot AI | 74.9 | 262.14K | $1.90 | $8.00 | — |
| Kimi K2.6 Moonshot AI | 74.6 | 262.14K | $0.95 | $4.00 | — |
| Kimi K2 Thinking Moonshot AI | 58.7 | 262.14K | $0.60 | $2.50 | — |
| Kimi K2 Thinking Turbo Moonshot AI | 58.7 | 262.14K | $1.15 | $8.00 | — |
| Kimi K2 Instruct 0905 Moonshot AI | — | 262.14K | — | — | — |
| Kimi K2 Instruct Moonshot AI | — | 131.07K | — | — | — |
| Kimi K2.5 Moonshot AI | 74.0 | 262.14K | $0.60 | $3.00 | — |
1–10 of 10 models · select a column to sort
About this order: Recommended shows our editorial shortlist first, followed by recent releases. The spec score is a 0–100 figure SyncDev calculates from each vendor's published specifications — context window, reasoning support, input modalities, tool calling, maximum output and how recently the model shipped. It measures breadth of capability, not benchmark performance, so a higher-scoring model is not automatically the better choice for your task.