Kimi K1.5 vs MiniMax-M2: which model should you choose?
We compared Kimi K1.5 and MiniMax-M2 on price, context window and 12 benchmarks. Kimi K1.5 has the stronger average benchmark score; Kimi K1.5 is cheaper on a blended 3:1 basis.
Kimi K1.5
Kimi K1.5 is a multimodal model from Moonshot AI. It offers a 128,000-token context window, costs $5.19 per million input tokens and $21.20 per million output tokens, and averages 80.7 across the index's benchmark suite.
- Input
- $5.19 / 1M
- Output
- $21.20 / 1M
- Context
- 128,000
- Avg score
- 80.7
- Open weights
- Yes
MiniMax-M2
MiniMax-M2 is a multimodal model from MiniMax. It offers a 200,000-token context window, costs $5.19 per million input tokens and $22.73 per million output tokens, and averages 80.6 across the index's benchmark suite.
- Input
- $5.19 / 1M
- Output
- $22.73 / 1M
- Context
- 200,000
- Avg score
- 80.6
- Open weights
- Yes
Specification comparison
| Spec | Kimi K1.5 | MiniMax-M2 | Edge |
|---|---|---|---|
| Input price / 1M | $5.19 | $5.19 | Tie |
| Output price / 1M | $21.20 | $22.73 | A |
| Context window | 128,000 | 200,000 | B |
| Max output | 4,096 | 65,536 | B |
| Average score | 80.7 | 80.6 | A |
| Arena Elo | 1432.2 | 1423.8 | A |
| Open weights | Yes | Yes | Tie |
Benchmark comparison
Kimi K1.5 wins 6, MiniMax-M2 wins 6, and 0 benchmarks are tied. The table shows the modelled score for each model and the gap.
| Benchmark | Kimi K1.5 | MiniMax-M2 | Gap | Winner |
|---|---|---|---|---|
| MMLU-Pro | 82.7 | 81.8 | 0.9 | Kimi K1.5 |
| GPQA Diamond | 74.4 | 75.4 | 1.0 | MiniMax-M2 |
| SWE-bench Verified | 68.2 | 69.9 | 1.7 | MiniMax-M2 |
| HumanEval | 87.1 | 88.0 | 0.9 | MiniMax-M2 |
| MATH-500 | 85.3 | 82.3 | 3.0 | Kimi K1.5 |
| AIME 2025 | 80.8 | 78.1 | 2.7 | Kimi K1.5 |
| LiveCodeBench | 71.1 | 73.4 | 2.3 | MiniMax-M2 |
| IFEval | 87.9 | 86.1 | 1.8 | Kimi K1.5 |
| TruthfulQA | 84.7 | 83.5 | 1.2 | Kimi K1.5 |
| HellaSwag | 87.2 | 90.4 | 3.2 | MiniMax-M2 |
| MMMU | 77.8 | 78.0 | 0.2 | MiniMax-M2 |
| Arena Elo | 1432.2 | 1423.8 | 8.4 | Kimi K1.5 |
Price analysis
On a blended 3:1 input-to-output basis, Kimi K1.5 costs about $9.19 per million tokens and MiniMax-M2 costs about $9.57. Kimi K1.5 is roughly 4% cheaper than MiniMax-M2. At high volume, that gap compounds quickly, so price should carry real weight unless the quality difference is decisive.
MiniMax-M2 has the larger context window at 200,000 tokens, compared with 128,000 for Kimi K1.5. That matters for long documents, large codebases and multi-step agent traces.
Verdict
Across 12 shared benchmarks, Kimi K1.5 wins 6, MiniMax-M2 wins 6, and 0 are tied. On the index average, Kimi K1.5 leads MiniMax-M2 by 0.1 points (80.7 versus 80.6).
The price picture favours Kimi K1.5. The context picture favours MiniMax-M2. Those two dimensions often matter more than a few benchmark points once a model is good enough for the task.
If you need a single default, start with Kimi K1.5 for quality-sensitive work, and check whether its price holds at your expected volume before committing.
Choose Kimi K1.5 if
- You want the higher average benchmark score in this index.
- Unit cost matters and you want the cheaper blended price.
Choose MiniMax-M2 if
- You need the larger 200,000-token context window.
- You need the larger 65,536-token output limit.
FAQ
Is Kimi K1.5 better than MiniMax-M2?
On the index average, Kimi K1.5 leads MiniMax-M2 by 0.1 points (80.7 versus 80.6). The better choice depends on price, context and your task. Kimi K1.5 wins 6 benchmarks, MiniMax-M2 wins 6.
Which is cheaper, Kimi K1.5 or MiniMax-M2?
Kimi K1.5 is cheaper on a blended 3:1 basis at about $9.19 per million tokens, versus $9.57 for MiniMax-M2.
Which has the larger context window?
MiniMax-M2 has the larger context window at 200,000 tokens. Kimi K1.5 has 128,000 tokens.
Are these models open weights?
Kimi K1.5 is open weights; MiniMax-M2 is open weights.
Which should I choose for production?
If you need a single default, start with Kimi K1.5 for quality-sensitive work, and check whether its price holds at your expected volume before committing.