Qwen3.8-Max vs Qwen3.8 2.4T A95B
Qwen3.8-Max is the cheaper option on input; Qwen3.8-Max is cheaper on output. Figures below are per million tokens, from vendor pricing cited in our reporting.
| Qwen3.8-Max | Qwen3.8 2.4T A95B | |
|---|---|---|
| Input per 1M tokens | $2 | $2 |
| Output per 1M tokens | $6 | $6 |
| Cached input per 1M | — | $0.25 |
| Context window | 1M tokens | 1.048576M tokens |
| Parameters | 2.4T total / ~95B active | — |
| Licence | proprietary | open weights |
| Benchmark | CodeArena WebDev: 1691 | — |
| Vendor | Alibaba | Alibaba |
| Released | 22 Sept 2026 | — |
What the numbers say
- Input price gap: 0% ($2 vs $2).
- Blended three-to-one input/output rate: $12.00 for Qwen3.8-Max, $12.00 for Qwen3.8 2.4T A95B.
- Context: 1M tokens against 1.048576M tokens.
- Licence: Qwen3.8-Max is proprietary; Qwen3.8 2.4T A95B is open weights.
Sources
Qwen3.8-Max specs · Qwen3.8 2.4T A95B specs · Full price index