Kimi K2 Thinking
DeprecatedKimi K2 Thinking is Moonshot AI (Kimi)'s language model with a 262K context window and up to 236K output tokens, available from 9 providers, starting at $0.6 / 1M input and $1.20 / 1M output. A reasoning-focused variant of Kimi K2 that performs step-by-step thinking while dynamically invoking tools for agentic problem solving.
| Specifications | |
|---|---|
moonshot-kimi-k2-thinking | |
| Language | |
| Deprecated | |
| Moonshot AI (Kimi) | |
| 262K tokens | |
| 236K tokens | |
| PDFText | |
| Text | |
| default | |
| 1.06T | |
| 1,695 | |
| 90,507 | |
| 1,327,565 | |
| · 11 months ago | |
| Benchmarks | |
|---|---|
| Intelligence Index | #150 |
| Math Index | #13 |
| MMLU-Pro | #39 |
| GPQA | #121 |
| HLE | #123 |
| LiveCodeBench | #18 |
| IFBench | #77 |
| Time to First Token | #451 |
| SciCode | #133 |
| AIME 2025 | #13 |
| LCR | #128 |
| TerminalBench Hard | #102 |
| TAU2 | #47 |
| Output TPS | #62 |
Capabilities
Input2/5
Text✓
Image·
Audio·
Video·
PDF✓
Output1/5
Text✓
Image·
Audio·
Video·
Embedding·
Capabilities8/13
Reasoning✓
Adaptive Reasoning·
Function Calling✓
Parallel Function Calling✓
Structured Outputs✓
Native JSON Schema✓
Web Search✓
URL Context·
Computer Use·
Code Execution·
File Search·
Prompt Caching✓
Assistant Prefill✓
Pricing by Provider
US Dollar ($)
Per 1M tokens
| Provider | Standard | Flex | Fast | Batch | |||||
|---|---|---|---|---|---|---|---|---|---|
| Input $ / 1M | Output $ / 1M | Cache Read $ / 1M | Input $ / 1M | Output $ / 1M | Input $ / 1M | Output $ / 1M | Input $ / 1M | Output $ / 1M | |
| $0.6 | $2.50 | N/A | $0.3 | $1.25 | $1.05 | $4.38 | $0.3 | $1.25 | |
| $0.6 | $2.50 | N/A | — | — | — | — | — | — | |
| $0.6 | $2.50 | N/A | — | — | — | — | — | — | |
| $2.50 | $2.50 | N/A | — | — | — | — | — | — | |
| $0.6 | $2.50 | N/A | — | — | — | — | — | — | |
| $0.8 | $1.20 | N/A | — | — | — | — | — | — | |
| $0.6 | $2.50 | $0.06 | — | — | — | — | — | — | |
| $0.6 | $2.50 | $0.15 | — | — | — | — | — | — | |
| $0.6 | $2.50 | $0.15 | — | — | — | — | — | — | |
Cost Calculator
US Dollar ($)
Preset:
Cheapest Instances to Run It
Cloud GPU instances that can host Kimi K2 Thinking, ranked by cheapest on-demand price. The model needs about 2539 GB of GPU memory at FP16 precision (estimated from its parameter count), so treat the fit as guidance rather than a guarantee.
All clouds
FP16 (full precision)
US Dollar ($)
No GPU instances fit Kimi K2 Thinking at FP16 precision — try a lower precision such as INT4.
Versions
| Version | Released | Context | Input / 1M | Output / 1M | Status |
|---|---|---|---|---|---|
| Kimi K2 Thinking | 262K | $0.600 | $1.20 | Current | |
| Kimi K2 | 262K | $0.570 | $2.30 | Available |