Model comparison
Qwen3.7-Max vs Kimi K2.6
- Input
- $2.50 / 1M tokens
- Output
- $7.50 / 1M tokens
- Context
- 1M
- Released
- 2026-05
- Input
- $0.95 / 1M tokens
- Output
- $4.00 / 1M tokens
- Context
- 256K
- Released
- 2026-04-20
Qwen3.7-Max scores higher on 3 of the 4 benchmarks both models report (the other leads on 1); Kimi K2.6 is cheaper on input tokens ($1 vs $2.5 per 1M, 2.6×); Qwen3.7-Max has the larger context window (1M vs 256K); Kimi K2.6 ships open weights while Qwen3.7-Max is API-only.
Specs and pricing side by side
| Spec | Qwen3.7-Max | Kimi K2.6 |
|---|---|---|
| Maker | Alibaba (Qwen) | Moonshot AI (Kimi) |
| Released | 2026-05 | 2026-04-20 |
| Context window | 1M | 256K |
| Max output | 65,536 tokens | 256K |
| Input price | $2.50 / 1M tokens | $0.95 / 1M tokens |
| Output price | $7.50 / 1M tokens | $4.00 / 1M tokens |
| Parameters | Not disclosed | 1T total (MoE), 32B active per token |
| License | Proprietary (closed weights); API-only via Alibaba Cloud Model Studio | Modified MIT (open weights; "Kimi K2" UI attribution required above 100M MAU or $20M monthly revenue) |
| Weights | API only | Open weights |
| Modalities | Text | Text, Vision, Video |
| Knowledge cutoff | Not disclosed | not published |
Benchmarks both models report
| Benchmark | Qwen3.7-Max | Kimi K2.6 |
|---|---|---|
| SWE-Bench Verified | 80.4% | 80.2% |
| SWE-Bench Pro | 60.6% | 58.6% |
| GPQA Diamond | 92.4% | 90.5% |
| Humanity's Last Exam | 41.4% | 54% |
Scores are as published by each maker or its cited source; harness, reasoning effort and date can differ between rows, so treat small gaps as ties. Sources are listed on each model page.
Which one should you pick?
Pick Qwen3.7-Max if…
- You care about SWE-Bench Verified, SWE-Bench Pro, GPQA Diamond (it scores higher on 3 shared benchmarks).
- You need the longer context (1M vs 256K).
- Your use case is: autonomous coding agents: multi-file refactors, debugging, and long-running build/test loops.
Pick Kimi K2.6 if…
- You care about Humanity's Last Exam (it scores higher on 1 shared benchmark).
- Cost matters: input tokens cost 2.6× less ($1 vs $2.5 per 1M).
- You want to self-host or fine-tune (open weights, Modified MIT (open weights; "Kimi K2" UI attribution required above 100M MAU or $20M monthly revenue)).
- You need Vision / Video input, which Qwen3.7-Max lacks.
- Your use case is: autonomous, multi-step software engineering and bug-fixing over large repositories.
FAQ
Which is better, Qwen3.7-Max or Kimi K2.6?
Qwen3.7-Max scores higher on 3 of the 4 benchmarks both models report (the other leads on 1); Kimi K2.6 is cheaper on input tokens ($1 vs $2.5 per 1M, 2.6×); Qwen3.7-Max has the larger context window (1M vs 256K); Kimi K2.6 ships open weights while Qwen3.7-Max is API-only. The better pick depends on whether you value benchmark scores, price per token, context length or openness for your workload.
How much cheaper is Kimi K2.6?
Qwen3.7-Max: $2.50 / 1M tokens input, $7.50 / 1M tokens output. Kimi K2.6: $0.95 / 1M tokens input, $4.00 / 1M tokens output (list API prices as published by each maker).
What are the context windows of Qwen3.7-Max and Kimi K2.6?
Qwen3.7-Max offers 1M with up to 65,536 tokens output; Kimi K2.6 offers 256K with up to 256K output.
Is Qwen3.7-Max newer than Kimi K2.6?
Qwen3.7-Max was released 2026-05; Kimi K2.6 was released 2026-04-20.