Alibaba · Available now
Qwen 3.8 (2026): every version, its price, and which ones you can try
Qwen3.8 is Alibaba’s model family: the Max version launched on August 3, 2026 (2.4 trillion parameters, 1M-token context), open weights followed on August 12 and 14, and OpenRouter lists six versions from $0.15 / $0.47 to $4 / $12 per million tokens; the chat on this page runs Qwen3.8 Max Prime, and Qwen 4 is still only a preview.
Use it now
Chat
≈ 4 credits per reply · See all models
New accounts get 5 free credits once, valid for 7 days. They work on the lower-priced models here; the priciest ones need credits you bought.
Start chatting — 5 free credits- Maker
- Alibaba
- Type
- Text
- Released
- 2026-08-03
- OpenRouter id
- qwen/qwen3.8-max-prime
Per million tokens
Price
| - | Input | Output | Source |
|---|---|---|---|
| Via OpenRouter (qwen/qwen3.8-max-prime) | $4 | $12 | openrouter.ai |
Checked
Facts, with sources
- Alibaba announced Qwen3.8-Max on August 3, 2026, available through the Alibaba Cloud Model Studio API; it described it as the most powerful model in its Qwen series to date.Source: alibabagroup.com · checked 2026-11-14
- Qwen3.8-Max is a sparse mixture-of-experts model with 2.4 trillion total parameters, about 95 billion active per inference, and a context window of up to 1 million tokens.Source: alibabagroup.com · checked 2026-11-14
- Open weights: Qwen3.8-2.4T-A95B went up on Hugging Face and ModelScope on August 12, 2026, and Qwen3.8-27B on August 14, 2026.Source: github.com · checked 2026-11-14
- Qwen3.8-27B is a dense vision-language model that reads images and video; its Hugging Face page lists the Apache 2.0 licence.Source: huggingface.co · checked 2026-11-14
- Qwen3.8-Flash-Next is an open-weight model that its Hugging Face card calls an experimental preview of the architecture that will underpin Qwen4; the card says Qwen3.8-Flash is the official version based on it, with 1M context by default.Source: huggingface.co · checked 2026-11-14
- Alibaba’s international list prices per million input / output tokens: qwen3.8-max $2 / $6, qwen3.8-flash $0.15 / $0.47, qwen3.8-27b $0.50 / $3.Source: docs.qwencloud.com · checked 2026-11-14
- OpenRouter lists six Qwen3.8 text models (per million input / output tokens): qwen/qwen3.8-max-prime $4 / $12 with 1M context; qwen/qwen3.8-max-0902 $2 / $6; qwen/qwen3.8-2.4t-a95b $2 / $6; qwen/qwen3.8-flash $0.15 / $0.47; qwen/qwen3.8-omni-flash $0.15 / $0.47; qwen/qwen3.8-27b from $0.035 / $1.05 depending on the host.Source: openrouter.ai · checked 2026-11-14
- OpenRouter describes Qwen3.8 Max Prime as a higher-throughput variant of Qwen3.8 Max, served as a separate SKU at a higher price; it takes text, image and video input.Source: openrouter.ai · checked 2026-11-14
- Qwen4-Max is now listed with a price: Alibaba Cloud’s Model Studio page for qwen4-max (updated November 16, 2026) calls it "the next-generation flagship multimodal model in the Qwen family". The earlier qwen4-max-preview ran through Token Plan only, until November 25, 2026 (UTC+8).Source: alibabacloud.com · checked 2026-11-16
Max versions
Qwen 3.8 Max: the three Max versions on OpenRouter
- OpenRouter lists the Max model three ways: Qwen3.8 Max Prime (a higher-throughput variant at $4 / $12), Qwen3.8 Max (0902) (an updated snapshot at $2 / $6) and Qwen3.8 2.4T A95B (the open-weight variant at $2 / $6), all priced per million input / output tokens (checked November 18, 2026).Source: openrouter.ai · checked 2026-11-18 · second-hand
Max versions
Qwen 3.8 Max Prime, Max (0902) and 2.4T-A95B compared
Searching for "qwen 3.8 max" lands on three different listings, because OpenRouter sells the Max model as separate entries. Max Prime is the pricier, higher-throughput one that the chat on this page runs; Max (0902) is a cheaper snapshot; 2.4T-A95B is the open-weight variant you can download. Alibaba’s own international list puts qwen3.8-max at $2 / $6 per million input / output tokens. For a newer Alibaba flagship see Qwen 4; other models from Chinese labs you can chat with here include Kimi K3 and DeepSeek V4.1 Flash.
| What | Max Prime | Max (0902) | 2.4T-A95B |
|---|---|---|---|
| OpenRouter id | qwen/qwen3.8-max-prime | qwen/qwen3.8-max-0902 | qwen/qwen3.8-2.4t-a95b |
| What OpenRouter says | A higher-throughput variant of Max, served as a separate SKU at a higher price | An updated snapshot of Qwen3.8 Max | The open-weight variant of Qwen3.8 Max, 95 billion active out of 2.4 trillion parameters |
| Price per million tokens (input / output) | $4 / $12 | $2 / $6 | $2 / $6 |
| Context window | 1,000,000 tokens | 1,000,000 tokens | 1,048,576 tokens |
| Input | Text, image and video | Text, image and video | Text |
| Listed on OpenRouter | September 23, 2026 | September 3, 2026 | August 12, 2026 |
| Typical reply at this price (our arithmetic) | about 4 credits | about 2 credits | about 2 credits |
Qwen3.8 27B
Qwen3.8 27B next to Qwen3.8 Flash
- Qwen3.8-27B is an open-weight dense vision-language model with 27B parameters under the Apache 2.0 licence; its model card gives 262,144 tokens of context natively, extensible to 1,000,000. OpenRouter lists 25 hosts for it, with input from $0.035 to $0.99 and output from $1.05 to $3.20 per million tokens (checked November 18, 2026).Source: openrouter.ai · checked 2026-11-18
Side by side
Qwen3.8 27B and Qwen3.8 Flash compared
People who search for Qwen3.8 27B are mostly weighing the open-weight 27B against the cheaper hosted Flash. The 27B has a very wide price range because 25 hosts on OpenRouter each set their own price, so the cheapest host and the dearest differ by about 28 times on input. The chat on this page runs Qwen3.8 Max Prime, not the 27B; for another open-weight model you can chat with here, see Inkling 2 Small. SenseTime’s 9B SenseNova Loop 9B uses the 27B as its yardstick.
| What | Qwen3.8 27B | Qwen3.8 Flash |
|---|---|---|
| What it is | Open-weight dense vision-language model, 27B parameters, Apache 2.0 licence | A multimodal reasoning model from Alibaba, for coding, agent work and long-video analysis |
| Context window | 262,144 tokens natively, extensible to 1,000,000 | 1,000,000 tokens |
| Input | Text, image and video | Text, image and video |
| Alibaba list price per million tokens (input / output) | $0.50 / $3 | $0.15 / $0.47 |
| OpenRouter hosts and price per million tokens | 25 hosts; input $0.035 to $0.99, output $1.05 to $3.20 | 2 hosts; $0.15 / $0.47 at both |
| Listed on OpenRouter | August 14, 2026 | August 26, 2026 |
Source: huggingface.coSource: openrouter.aiSource: docs.qwencloud.com
Nearest available
Models you can use now instead
Qwen3.8-Max
Alibaba$2 / $6 (per million tokens)
- Alibaba’s price for qwen3.8-max is $2 / $6 per million input / output tokens (Singapore region).Source: docs.qwencloud.com · checked 2026-11-13
- Alibaba describes qwen3.8-max as a 2.4-trillion-parameter mixture-of-experts model.Source: docs.qwencloud.com · checked 2026-11-13
- Listed on OpenRouter as qwen/qwen3.8-max-0902 at $2 / $6 per million tokens.Source: openrouter.ai · checked 2026-11-13
Qwen3.8 Flash
Alibaba$0.15 / $0.47 (per million tokens)
- Alibaba’s price for qwen3.8-flash is $0.15 / $0.47 per million input / output tokens (Singapore, EU, Hong Kong and US regions).Source: docs.qwencloud.com · checked 2026-11-13
- Listed on OpenRouter as qwen/qwen3.8-flash at $0.15 / $0.47 per million tokens.Source: openrouter.ai · checked 2026-11-13
FAQ
Questions
What is Qwen 3.8?
Alibaba’s Qwen model generation of 2026. Alibaba announced the Max version on August 3, 2026 as the most powerful model in the Qwen series so far, followed by open-weight models on Hugging Face and ModelScope.
Which Qwen 3.8 versions are there?
On OpenRouter: Max Prime, Max (0902), 2.4T-A95B, Flash, Omni Flash and 27B. Alibaba also has Qwen3.8-Flash-Next, an open-weight preview of the architecture it says will underpin Qwen4.
How much does Qwen 3.8 cost?
Per million input / output tokens on OpenRouter: Max Prime $4 / $12, Max (0902) and 2.4T-A95B $2 / $6, Flash and Omni Flash $0.15 / $0.47, 27B from $0.035 / $1.05 depending on the host (25 hosts). Alibaba’s own international list has Max at $2 / $6 and Flash at $0.15 / $0.47.
Which Qwen 3.8 model can I try here?
The chat on this page runs Qwen3.8 Max Prime. You can also pick Qwen3.8-Max or Qwen3.8 Flash in the chat, on credits and without an Alibaba account.
Can I download Qwen 3.8, and what is Qwen 27B (qwen 27b)?
Some versions, yes. Qwen3.8-2.4T-A95B (August 12, 2026) and Qwen3.8-27B (August 14, 2026) are on Hugging Face and ModelScope; the 27B page lists the Apache 2.0 licence. Qwen3.8-Max itself is reached through the API. "Qwen 27B" means Qwen3.8-27B, the open-weight 27B-parameter dense vision-language model (OpenRouter id qwen/qwen3.8-27b, 25 hosts, from $0.035 per million input tokens; see the 27B table on this page); OpenRouter also still lists the older qwen3.6-27b and qwen3.5-27b, which cost more per token. The chat here runs Max Prime, not the 27B.
What is Qwen 3.8 Max, and how is Max Prime different?
Alibaba describes Qwen3.8 Max as a sparse mixture-of-experts model with 2.4 trillion total parameters, about 95 billion active per inference, and a context window of up to 1 million tokens. OpenRouter sells it as Max Prime ($4 / $12 per million tokens, a higher-throughput variant), Max (0902) ($2 / $6, an updated snapshot) and 2.4T-A95B ($2 / $6, open weights); the table on this page compares them.
How does Qwen 3.8 relate to Qwen 4?
Qwen 4 is the next generation: Alibaba now lists Qwen4-Max with a price on Model Studio, after a time-limited preview (qwen4-max-preview, through Token Plan until November 25, 2026). We found no Qwen4 open weights or smaller Qwen4 tiers; see Qwen 4 for what is confirmed and what is rumoured.
Is this an Alibaba page?
No. This page is unofficial and not affiliated with or endorsed by Alibaba. Every fact links to its source.