Model comparison
Qwen3.6 vs Qwen3.5
- Input
- $0.2885 / 1M tokens
- Output
- $3.17 / 1M tokens
- Context
- 262K (extensible ~1M)
- Released
- 2026-04
- Input
- $0.54 / 1M tokens
- Output
- $3.40 / 1M tokens
- Context
- 256K
- Released
- 2026-02-16
Qwen3.6 and Qwen3.5 split the 6 benchmarks both report evenly; Qwen3.6 is cheaper on input tokens ($0.3 vs $0.5 per 1M, 1.9×); Qwen3.6 has the larger context window (262K vs 256K).
Specs and pricing side by side
| Spec | Qwen3.6 | Qwen3.5 |
|---|---|---|
| Maker | Alibaba (Qwen) | Alibaba (Qwen)same |
| Released | 2026-04 | 2026-02-16 |
| Context window | 262K (extensible ~1M) | 256K |
| Max output | ~32K (up to ~82K for complex reasoning) | not published |
| Input price | $0.2885 / 1M tokens | $0.54 / 1M tokens |
| Output price | $3.17 / 1M tokens | $3.40 / 1M tokens |
| Parameters | 27B dense (Qwen3.6-27B); 35B total / 3B active MoE (Qwen3.6-35B-A3B) | 397B total / 17B active (MoE, 512 experts) |
| License | Apache 2.0 | Apache-2.0 |
| Weights | Open weights | Open weightssame |
| Modalities | Text, Vision, Video | Text, Vision, Video, PDF |
| Knowledge cutoff | Not publicly disclosed | not published |
Benchmarks both models report
| Benchmark | Qwen3.6 | Qwen3.5 |
|---|---|---|
| SWE-bench Verified | 77.2% | 76.4% |
| Terminal-Bench 2.0 | 59.3% | 52.5% |
| MMLU-Pro | 86.2% | 87.8% |
| AIME 2026 | 94.1% | 91.3% |
| GPQA Diamond | 87.8% | 88.4% |
| MMMU Pro | 75.8% | 79% |
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 Qwen3.6 if…
- You care about SWE-bench Verified, Terminal-Bench 2.0, AIME 2026 (it scores higher on 3 shared benchmarks).
- Cost matters: input tokens cost 1.9× less ($0.3 vs $0.5 per 1M).
- You need the longer context (262K vs 256K).
- Your use case is: agentic coding and repository-level software engineering (SWE-bench / Terminal-Bench style tasks).
Pick Qwen3.5 if…
- You care about MMLU-Pro, GPQA Diamond, MMMU Pro (it scores higher on 3 shared benchmarks).
- You need PDF input, which Qwen3.6 lacks.
- Your use case is: self-hosted agentic systems that need open weights for tool use, browsing, and terminal/coding workflows.
FAQ
Which is better, Qwen3.6 or Qwen3.5?
Qwen3.6 and Qwen3.5 split the 6 benchmarks both report evenly; Qwen3.6 is cheaper on input tokens ($0.3 vs $0.5 per 1M, 1.9×); Qwen3.6 has the larger context window (262K vs 256K). The better pick depends on whether you value benchmark scores, price per token, context length or openness for your workload.
How much cheaper is Qwen3.6?
Qwen3.6: $0.2885 / 1M tokens input, $3.17 / 1M tokens output. Qwen3.5: $0.54 / 1M tokens input, $3.40 / 1M tokens output (list API prices as published by each maker).
What are the context windows of Qwen3.6 and Qwen3.5?
Qwen3.6 offers 262K (extensible ~1M) with up to ~32K (up to ~82K for complex reasoning) output; Qwen3.5 offers 256K.
Is Qwen3.6 newer than Qwen3.5?
Qwen3.6 was released 2026-04; Qwen3.5 was released 2026-02-16.