Kimi K2: pricing, context window and benchmarks
Kimi K2 is a multimodal model from Moonshot AI. It offers a 128,000-token context window, costs $5.85 per million input tokens and $21.72 per million output tokens, and averages 83.0 across the index's benchmark suite.
Pricing
Kimi K2 costs $5.85 per million input tokens and $21.72 per million output tokens. A blended 3:1 input-to-output mix works out to about $9.82 per million tokens. Cached input is listed at $0.62 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $5.85 | per 1M tokens |
| Output | $21.72 | per 1M tokens |
| Cached input | $0.62 | per 1M tokens |
| Blended 3:1 | $9.82 | estimate per 1M tokens |
| 1M in + 1M out | $27.57 | estimate |
Benchmarks
Scores below are modelled estimates on a 0–100 scale. Use them to compare relative position, not as official vendor results.
| Benchmark | Score | Rank | Percentile | Relative |
|---|---|---|---|---|
| MMLU-Pro | 83.3 | 99 | 84th | |
| GPQA Diamond | 77.5 | 97 | 84th | |
| SWE-bench Verified | 70.2 | 105 | 83th | |
| HumanEval | 87.1 | 85 | 86th | |
| MATH-500 | 85.3 | 104 | 83th | |
| AIME 2025 | 85.1 | 38 | 94th | |
| LiveCodeBench | 78.5 | 61 | 90th | |
| IFEval | 91.5 | 44 | 93th | |
| TruthfulQA | 84.8 | 67 | 89th | |
| HellaSwag | 92.3 | 38 | 94th | |
| MMMU | 77.5 | 107 | 82th | |
| Arena Elo | 1458.5 | 56 | 91th |
What Kimi K2 is good for
Kimi K2 is tuned toward coding work. It is a reasonable candidate for code generation, review, debugging and repository-scale tasks, especially when paired with a retrieval layer. At $5.85 per million input tokens, it sits in the mid-range of the index, and the 128,000-token context window is enough for large files and multi-file diffs.
Strengths
- Large 128,000-token context window for long documents and agent traces.
- Open weights make self-hosting, fine-tuning and data-residency review possible.
- Accepts more than text, including image input where the provider supports it.
- Reasoning-oriented profile, useful for multi-step problems and structured analysis.
- Coding benchmarks are a relative strength within its tier.
Trade-offs
- No standout weakness in the indexed specifications, but real-world behaviour still needs testing.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| DeepSeek-R1 Instruct | DeepSeek | 1000000 | $5.90 | 83.1 |
| Grok 3 Turbo | xAI | 128000 | $15.97 | 83.1 |
| Kimi K2 | Moonshot AI | 128000 | $5.85 | 83.0 |
| GPT-4.1 Turbo | OpenAI | 1000000 | $16.12 | 82.9 |
| Grok 3 Reasoning | xAI | 200000 | $16.07 | 82.9 |
| Grok 3 | xAI | 128000 | $15.84 | 82.8 |
FAQ
How much does Kimi K2 cost?
Kimi K2 is listed at $5.85 per million input tokens and $21.72 per million output tokens. Cached input is $0.62 per million tokens. A blended 3:1 input-to-output mix is about $9.82 per million tokens.
What is the context window of Kimi K2?
Kimi K2 has a context window of 128,000 tokens and can emit up to 65,536 output tokens per request in this index.
Is Kimi K2 open weights?
Yes. Kimi K2 is listed as an open-weight model under a open weights licence.
How does Kimi K2 score on benchmarks?
Kimi K2 averages 83.0 across the scored benchmarks in this index, ranking #74 of 600 models. Its strongest relative result is Arena Elo at 1458.5.
What is Kimi K2 best for?
Kimi K2 is tuned toward coding work. It is a reasonable candidate for code generation, review, debugging and repository-scale tasks, especially when paired with a retrieval layer. At $5.85 per million input tokens, it sits in the mid-range of the index, and the 128,000-token context window is enough for large files and multi-file diffs.