Same open weights, served by 3 providers.
| Provider | Their model id | Serving | Input | Output | Cached in | Blended 3:1 | Source | |
|---|---|---|---|---|---|---|---|---|
DeepInfra Deprecated, 2026-09-07 | moonshotai/Kimi-K2-Thinking | standardFP4Reduced context (131k of 262k) | $0.47 | $2.0 | $0.141 | $0.853 | Verified 4h ago · Changed 7h ago | |
Azure AI Foundry | kimi-k2-thinking | standardglobal endpointQuantization not disclosed | $0.60 | $2.5 | — | $1.07 | Verified 4h ago · Changed 6h ago | |
AWS Bedrock | moonshot.kimi-k2-thinking | standardQuantization not disclosed | $0.60 | $2.5 | — | $1.07 | Verified 6h ago · Changed 6h ago | |
Azure AI Foundry | kimi-k2-thinking | standarddata zone (US)Quantization not disclosed | $0.66 | $2.75 | — | $1.18 | Verified 4h ago · Changed 6h ago |
Same weights don't always mean identical outputs: quantization, context limits and serving stacks can change quality, especially for tool calling. Prices are USD per 1M tokens; hyperscaler prices are for the reference region shown.
| DeepInfra fp4 | $87/mo |
| Azure AI Foundry global endpoint | $110/mo |
| AWS Bedrock | $110/mo |
| Azure AI Foundry data zone (US) | $121/mo |
Cache hits use each provider's cached-input price where it has one. Cache write costs aren't included.