macrostack

AI API price index · model maker · read 17 Nov 2026

Alibaba Cloud Model Studio (Qwen) API pricing

33 prices as Alibaba Cloud Model Studio (Qwen) publishes them — 33 text models — read on its own pricing page on 17 Nov 2026. Its current text models cost $0.13 to $7.50 per million output tokens.

Free tier A free quota of 1 million tokens on several models (among them qwen4-max, qwen3.8-max, qwen3.7-max, qwen3.7-plus, qwen3.8-flash, qwen3.7-flash, qwen3.8-27b), "Valid for 90 days from the date of Model Studio activation, model release, or application approval, whichever is later", and "only in Singapore. No free quota is available in other regions."

What to know before you build on Alibaba Cloud Model Studio (Qwen)

  • Prices here are the International (Singapore) ones: qwen4-max is $1.50 in, $0.075 cached and $4.50 out per million tokens.
  • Plus and Flash models are priced by prompt length, and the whole request is billed at its band's price: qwen3.6-plus is $0.50 in and $3.00 out up to 256K tokens, $2.00 and $6.00 above.
  • Batch calls cost 50% of the real-time price, and the batch and context-cache discounts do not apply together.
  • A free quota of 1 million tokens per model lasts 90 days, and only in the Singapore region.
  • The China (Beijing) region is cheaper: qwen4-max is $1.413 in and $4.24 out there, the price DeepInfra and Novita charge for it.
  • Open-weight Qwen models are sold here too, from $0.25 in and $2.00 out (qwen3.5-35b-a3b); qwen3.8-27b is $0.50 and $3.00.

Against the other 31

Across the index, 390 current text models run from $0.03 to $50.00 per million output tokens — 1667×. Alibaba Cloud Model Studio (Qwen)’s cheapest, qwen3.7-flash at $0.13, has 9 cheaper text models elsewhere; its dearest, qwen3.7-max, costs $7.50. A cheaper model is a different model, not the same one for less — compare capability before price.

The same model elsewhere

Open-weight models Alibaba Cloud Model Studio (Qwen) sells that other hosts price too — the only like-for-like comparison in the index.

  • Qwen3.8 27B — Cerebras $0.99 / $1.49 · DeepInfra $0.20 / $2.50 · Alibaba Cloud Model Studio (Qwen) $0.50 / $3.00 · Nebius Token Factory $0.45 / $3.00 · Novita AI $0.42 / $3.00 · Cloudflare Workers AI $0.45 / $3.20 · Groq $0.80 / $4.00 — 2.7× apart

Input / output per 1M tokens, standard tier.

Text models

ModelInputCached inputOutputNote
qwen4-max$1.50$0.075$4.500<Token≤1M; cache write $1.875 (5m) or $3 (1h) per 1M tokens
qwen3.8-max$2.00—$6.000<Token≤1M
qwen3.7-max$2.50—$7.500<Token≤1M
qwen-max$1.60—$6.40non-thinking mode only
qwen3.7-plus$0.40—$1.600<Token≤256K; list price, marked "Limited-time 20% off" (the discounted price is shown in the console, not on the page)
qwen3.7-plus$1.20—$4.80Long context: input of 256K<Token≤1M tokens per request (every token billed at this band's price); list price, marked "Limited-time 20% off" · Not ranked: long-prompt band; the index ranks the standard band.
qwen3.6-plus$0.50—$3.000<Token≤256K
qwen3.6-plus$2.00—$6.00Long context: input of 256K<Token≤1M tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen3.5-plus$0.40—$2.400<Token≤256K
qwen3.5-plus$0.50—$3.00Long context: input of 256K<Token≤1M tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen-plus$0.40—$1.200<Token≤256K; non-thinking output; thinking-mode output $4 per 1M tokens
qwen-plus$1.20—$3.60Long context: input of 256K<Token≤1M tokens per request (every token billed at this band's price); non-thinking output; thinking-mode output $12 per 1M tokens · Not ranked: long-prompt band; the index ranks the standard band.
qwen3.8-flash$0.15—$0.470<Token≤1M
qwen3.7-flash$0.03—$0.130<Token≤32K
qwen3.7-flash$0.10—$0.40Long context: input of 32K<Token≤256K tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen3.7-flash$0.20—$0.80Long context: input of 256K<Token≤1M tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen3.6-flash$0.25—$1.500<Token≤256K
qwen3.6-flash$1.00—$4.00Long context: input of 256K<Token≤1M tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen3.5-flash$0.10—$0.400<Token≤1M
qwen-flash$0.05—$0.400<Token≤256K
qwen-flash$0.25—$2.00Long context: input of 256K<Token≤1M tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen3-coder-flash$0.30—$1.500<Token≤32K
qwen3-coder-flash$0.50—$2.50Long context: input of 32K<Token≤128K tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen3-coder-flash$0.80—$4.00Long context: input of 128K<Token≤256K tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen3-coder-flash$1.60—$9.60Long context: input of 256K<Token≤1M tokens per request (every token billed at this band's price) · Not ranked: long-prompt band; the index ranks the standard band.
qwen3.8-2.4t-a95b$2.00—$6.00open-weight model; 0<Token≤1M
qwen3.8-27b$0.50—$3.00open-weight model; 0<Token≤1M
qwen3.6-35b-a3b$0.375—$2.25open-weight model; 0<Token≤256K
qwen3.6-27b$0.60—$3.60open-weight model; 0<Token≤256K
qwen3.5-397b-a17b$0.60—$3.60open-weight model; 0<Token≤256K
qwen3.5-122b-a10b$0.40—$3.20open-weight model; 0<Token≤256K
qwen3.5-27b$0.30—$2.40open-weight model; 0<Token≤256K
qwen3.5-35b-a3b$0.25—$2.00open-weight model; 0<Token≤256K

22 of these 33 prices are ranked in the index; the rest say why in their note. Text and embeddings per 1M tokens unless the row says otherwise.

Discounts and surcharges, as published

  • Batch calls: "the unit price for both input and output tokens is 50% of the real-time inference price" on models that support batch. Batch and context-cache discounts "cannot apply simultaneously".
  • Context cache: "explicit cache creation is billed at 125% of the standard input price, and cache hits at 10%". qwen4-max prints its own: cache write $1.875 (5 minutes) or $3 (1 hour), cache read $0.075 per 1M tokens.
  • qwen3.7-plus is marked "Limited-time 20% off" beside its list price; the page says it "only lists standard prices" and that current promotions are in the Model Studio console. The index ranks the list price.
  • Other regions have their own tables: China (Beijing) is cheaper (qwen4-max $1.413 in and $4.24 out), and Germany (Frankfurt), US (Virginia) and Japan (Tokyo) are listed too.

Where these numbers come from

Read on alibabacloud.com/help/en/model-studio/model-pricing on 17 Nov 2026. Macrostack measures nothing and is paid by no provider on this index. Prices change; the date is the claim.

The other providers in the index

Questions

How much does the Alibaba Cloud Model Studio (Qwen) API cost?
On 17 Nov 2026, Alibaba Cloud Model Studio (Qwen)'s current text models cost $0.13 to $7.50 per million output tokens at the standard tier (qwen3.7-flash to qwen3.7-max). The full list, with input and cached-input prices, is on this page.
Does Alibaba Cloud Model Studio (Qwen) have a free tier?
What Alibaba Cloud Model Studio (Qwen)'s own pages say: A free quota of 1 million tokens on several models (among them qwen4-max, qwen3.8-max, qwen3.7-max, qwen3.7-plus, qwen3.8-flash, qwen3.7-flash, qwen3.8-27b), "Valid for 90 days from the date of Model Studio activation, model release, or application approval, whichever is later", and "only in Singapore. No free quota is available in other regions."
Does Alibaba Cloud Model Studio (Qwen) offer discounts?
Batch calls: "the unit price for both input and output tokens is 50% of the real-time inference price" on models that support batch. Batch and context-cache discounts "cannot apply simultaneously". Context cache: "explicit cache creation is billed at 125% of the standard input price, and cache hits at 10%". qwen4-max prints its own: cache write $1.875 (5 minutes) or $3 (1 hour), cache read $0.075 per 1M tokens. qwen3.7-plus is marked "Limited-time 20% off" beside its list price; the page says it "only lists standard prices" and that current promotions are in the Model Studio console. The index ranks the list price. Other regions have their own tables: China (Beijing) is cheaper (qwen4-max $1.413 in and $4.24 out), and Germany (Frankfurt), US (Virginia) and Japan (Tokyo) are listed too.
The Macrostack brief

New swaps, worth your inbox.

A short, occasional email when we add a high-intent alternative or ship a new head-to-head. No spam, no selling your address — unsubscribe in one click.