A side-by-side comparison of GPT-5.6 Luna and Kimi K2.8 Preview: input/output pricing, capabilities and available endpoints, served live from J8API. Both are reachable with the same API key.
| Item | GPT-5.6 Luna | Kimi K2.8 Preview |
|---|---|---|
| Input (per 1M tokens) | $0.13 | $0.55 |
| Output (per 1M tokens) | $0.78 | $2.2 |
| Cache read explicit (per 1M tokens) | $0.013 | $0.1375 |
| Cache write 5m (per 1M tokens) | $0.1625 | — |
| Item | GPT-5.6 Luna | Kimi K2.8 Preview |
|---|---|---|
| Context window | 922,000 | 1,048,576 |
| Max output | 128,000 | 131,072 |
| Item | GPT-5.6 Luna | Kimi K2.8 Preview |
|---|---|---|
| function_calling | Yes | Yes |
| prompt_caching | Yes | Yes |
| reasoning | — | Yes |
| vision | Yes | — |
| web_search | — | Yes |
On input, GPT-5.6 Luna is cheaper ($0.13 vs $0.55 per 1M tokens). On output, GPT-5.6 Luna is cheaper ($0.78 vs $2.2 per 1M tokens). All prices are per million tokens in USD.
Kimi K2.8 Preview does — GPT-5.6 Luna accepts 922,000 input tokens and Kimi K2.8 Preview accepts 1,048,576.
both support function_calling, prompt_caching; only GPT-5.6 Luna supports vision; only Kimi K2.8 Preview supports reasoning, web_search.
Yes. Both are available on J8API through one API key and the same OpenAI-compatible endpoint — switching means changing the model field from "gpt-5.6-luna" to "kimi-k2.8-preview", nothing else.
Both are available on J8API under one API key — switching between them means changing the model field and nothing else, so you can use each where it fits rather than picking one.