One agent session
1M input tokens and 100K output tokens, no caching.
- Input1M x $2 = $2.00
- Output0.1M x $6 = $0.60
$2.60per session
Alibaba Cloud (Qwen) - Closed weights
Qwen3.8-Max is a closed-weights model from Alibaba Cloud (Qwen) that costs $2 per 1M input tokens and $6 per 1M output tokens on the first-party API, with a 1M-token context window.
Last verified from alibabacloud.com. All sources
| Item | Price / 1M tokens |
|---|---|
| Input | $2 |
| Output | $6 |
| Cache read | $0.17 |
| Cache write (5 minutes) | $2.50 |
Note: Prices are the International (Singapore) deployment list prices; the Global deployment scope is priced lower on the same page ($1.65 / $4.951). The text generation page lists it as a recommended model. Thinking and non-thinking modes cost the same. cache read/cache write are explicit-cache read and creation (an explicit cache lasts 5 minutes and resets on a hit); implicit cache hits are $0.25 per 1M. The context-cache doc says explicit-cache hits on qwen3.8-max are not the usual 10% of input. The Singapore pricing row carries no batch discount, so batch prices are blank. Snapshot qwen3.8-max-0902 costs the same. The console states context as 1M and max output as 128K; recorded as 1,000,000 (the pricing page defines 1M = 1,000,000 tokens) and 131,072 (the QwenCloud model page states 131K).
qwen3.8-maxalibaba/qwen3.8-maxJSON Schema mode is documented for this model family, but the structured output doc says Singapore region models are not supported yet; batch is listed only for the China (Beijing) region, not Singapore. Not stated means the provider's docs do not say either way. It is not the same as No. Checked against provider docs on . Compare capabilities across models.
Worked from the standard list rates above, with the math shown so you can swap in your own token counts.
1M input tokens and 100K output tokens, no caching.
$2.60per session
10,000 requests a day at 2K input and 500 output tokens each.
$70.00per day
$2,100.00 per 30-day month
90% of the input is read from the prompt cache. The one-time cache write premium is not included.
$0.953per session
| Model | Input / 1M | Output / 1M |
|---|---|---|
| Qwen3.8-Max (this page) | $2 | $6 |
| Grok 4.5xAI | $2 | $6 |
| Grok 4.6xAI | $2 | $6 |
| Grok 4.7xAI | $2 | $6 |
| Qwen3.8-2.4T-A95BAlibaba Cloud (Qwen) | $2 | $6 |
| Qwen-MaxAlibaba Cloud (Qwen) | $1.60 | $6.40 |
| Qwen3.7-MaxAlibaba Cloud (Qwen) | $2.50 | $7.50 |
| Qwen3.5-Omni-PlusAlibaba Cloud (Qwen) | $1.40 | $8.30 |
| Qwen3.5-397B-A17BAlibaba Cloud (Qwen) | $0.60 | $3.60 |
| Qwen3.6-27BAlibaba Cloud (Qwen) | $0.60 | $3.60 |
Side-by-side pages with both models' prices, specs, cost examples and like-for-like benchmark rows.
Qwen3.8-Max costs $2 per 1M input tokens and $6 per 1M output tokens on Alibaba Cloud (Qwen)'s first-party API, per alibabacloud.com on October 26, 2026. Per 1M tokens, cache reads cost $0.17.
Qwen3.8-Max has a 1,000,000-token context window and returns up to 131,072 output tokens per response.
No. Qwen3.8-Max is a closed-weights model. You call it through Alibaba Cloud (Qwen)'s API as qwen3.8-max or through Vercel AI Gateway as alibaba/qwen3.8-max.
Alibaba Cloud (Qwen) released Qwen3.8-Max on August 2, 2026.
Last verified from alibabacloud.com.
Newsletter
One email per week. New model releases, price cuts and what they mean for your bill.