Qwen 3.8 Max
Alibaba · released 2 Sep 2026 · proprietary API
- Newest release
Alibaba's current Max tier, 20% cheaper on both halves than Qwen 3.7 Max; thinking and non-thinking modes on one model ID.
Context window
1M
1,000,000 tokens
Input price
$2
per 1M tokens
Output price
$6
per 1M tokens
Blended
$3
3:1 in:out mix
Weights
Proprietary
API only
Released
2 Sep 2026
Alibaba
Long-context / tier pricing — Cache hits bill at 10% of input ($0.20/1M); explicit cache creation at 125%. Batch calls are half price on both halves, and cannot be combined with the cache discount. Listed prices are the base rate for prompts up to ~200K tokens.
Where it ranks
Of 25 models tracked.
Cost to run
#10(tied)
of 25
Input price
#12(tied)
of 25
Context window
#5(tied)
of 25
Compare with
The nearest models by blended cost to run.
Put it to work
Specs from BitByteCore's hand-verified Frontier Ledger (v2026-09-03), this row verified 3 Sep 2026 against Alibaba's official pricing ↗. Model pricing moves fast, so confirm with Alibaba before you rely on a figure.