Alibaba Qwen models: the cheapest way to a million-token context

Prices from official vendor price lists · October 2026
Alibaba has 34 models in this rating, led by Qwen3.8 Max Prime and Qwen3.8 Max. 39 of them have a public per-token price, from $0.13 to $10.08 per 1M output tokens, and context windows reach 1.05M. 32 have a free tier; 32 ship open weights.
Alibaba (Qwen)
Services found: 34
#AI modelCapabilitiesAPI $/1M (in / out)Context windowPrice / moOutput $/1MOverviewVisit site
21

Qwen3 VL 235B A22B Instruct

Qwen
Text & chatVisionReasoning
$0.21 / $1.9262.1K tokensAPI only1.9View
22

Qwen3.6 Plus

Qwen
Text & chatVisionVideo
$0.325 / $1.951M tokensAPI only1.95View
23

Qwen3 VL 8B Instruct

Qwen
Text & chatVisionReasoning
$0.117 / $0.455262.1K tokensAPI only0.455View
24

Qwen3.6 Flash

Qwen
Text & chatVisionVideo
$0.1875 / $1.1251M tokensAPI only1.125View
25

Qwen3.5-27B

Qwen
Text & chatVisionVideo
$0.195 / $1.56262.1K tokensAPI only1.56View
26

Qwen3 Next 80B A3B Instruct

Qwen
Text & chatReasoningAdjustable thinking
$0.09 / $1.1262.1K tokensAPI only1.1View
27

Qwen3 VL 30B A3B Instruct

Qwen
Text & chatVisionReasoning
$0.15 / $0.6262.1K tokensAPI only0.6View
28

Qwen3 Max

Qwen
Text & chatReasoningAdjustable thinking
$0.78 / $3.9262.1K tokensAPI only3.9View
29

Qwen3 Coder Flash

Qwen
Text & chatVisionCoding
$0.195 / $0.9751M tokensAPI only0.975View
30

Qwen3.5 Plus 2026-02-15

Qwen
Text & chatVisionVideo
$0.26 / $1.561M tokensAPI only1.56View
31

Qwen3.5 Plus 2026-04-20

Qwen
Text & chatVisionVideo
$0.3 / $1.81M tokensAPI only1.8View
32

Qwen3 Coder Plus

Qwen
Text & chatCodingReasoning
$0.65 / $3.251M tokensAPI only3.25View
33

Qwen3 Max Thinking

Qwen
Text & chatReasoningAdjustable thinking
$0.78 / $3.9262.1K tokensAPI only3.9View
34

Qwen3.8 Omni Flash

Qwen
Text & chatVisionAudio
1M tokensAPI onlyView

Browse by need

Reasoning models169 modelsCoding assistants17 modelsVision and images156 modelsLong context 1M+82 modelsFree tier available80 modelsOpen weights88 models

Alibaba ships more models into this rating than anyone else: 34 entries spanning the Qwen3.8 and Qwen3.7 generations, coder variants, omni models and the Qwen-Long family. Most carry open weights, and pricing starts at $0.028 input per 1M tokens — low enough that cost stops being the deciding factor.

The practical draw is range. Qwen3.8-Max handles frontier-adjacent work, Qwen3.8-Flash covers high-volume routing for cents, and Qwen-Long stretches to a 10M-token window that no other vendor matches. Model Studio also grants a free allowance per model on a 90-day window, so evaluation costs nothing.

Widest line-up34 models, from a 10M-token long-context option to sub-cent flash variants.
Open weights by defaultMost of the Qwen family is downloadable, so self-hosting stays on the table.
Free to evaluateModel Studio grants roughly a million tokens per model before you pay anything.

Models from this vendor

34 Alibaba (Qwen) models in this rating. Prices are input and output per 1M tokens from the vendor official price list; for dirhams multiply by 3.6725. Tap a card to open the model page.

Qwen3.8 Max Prime
#1
Qwen
in $3.36out $10.081M tokensText & chatReasoning
Qwen3.8 Max (0902)
#2
Alibaba (Qwen)
in $2out $61M tokensText & chatVision
Qwen3.8 2.4T A95B
#3
Alibaba (Qwen)
in $2out $61.05M tokensText & chatReasoning
Qwen3.8 27B
#4
Alibaba (Qwen)
in $0.42out $31M tokensText & chatVision
Qwen3.7 Max
#5
Alibaba (Qwen)
in $2.5out $7.51M tokensText & chatReasoning
Qwen3.7 Plus
#6
Alibaba (Qwen)
in $0.32out $1.281M tokensText & chatVision
Qwen3.6 27B
#7
Alibaba (Qwen)
in $0.3out $2262.1K tokensText & chatVision
Qwen3.5 397B A17B
#8
Alibaba (Qwen)
in $0.55out $3.5262.1K tokensText & chatVision

How to choose

  • Use the free allowance first — roughly a million tokens per model on a 90-day window.
  • Match the variant to the job: Flash for volume, Plus for balance, Max for quality, Qwen-Long for archives.
  • Check region: Singapore and international rate cards differ on the same model.
  • Consider self-hosting once volume is steady, since most weights are downloadable.

Common mistakes

Treating all Qwen variants as equivalentThe spread between Flash and Max is more than fiftyfold on output.
Missing the busy-hour pricing on resold modelsDeepSeek entries on Model Studio carry peak and off-peak rates.

Who this suits

High-volume operationsThe cheapest credible models in the rating, at sub-cent rates.
Long-document workQwen-Long 10M window has no equivalent elsewhere.
Self-hostersOpen weights across most of the family.

Related filters

Text & chat228 models for text generation and chat
Vision156 models that read images
Reasoning169 models that think before answering
Adjustable thinking79 models with a reasoning depth dial
OpenAI30 models in this rating
Google DeepMind17 models in this rating
Anthropic13 models in this rating

Frequently asked questions

Is Qwen free to use?

Model Studio includes a free allowance per model on a 90-day window, and most Qwen weights are downloadable for self-hosting.

Which Qwen model should I start with?

Qwen3.8-Flash for volume, Qwen3.7-Plus for balanced work, Qwen3.8-Max when quality matters most.

How does Qwen handle Arabic?

It is trained with heavier non-English weighting than the US labs, which shows on multilingual tasks. Test on your own text.

Is Qwen open source?

Most of the family publishes weights; licences vary by model, so check the specific release.

What is the difference between Qwen3.7 and 3.8?

3.8 is the current generation; 3.7 remains available and cheaper.

Frequently searched

By capabilityReasoningCodingVisionLong context 1M+Free tierOpen weightsLow latencyRealtime voiceImage generation
By vendorAnthropic (13)OpenAI (30)Google DeepMind (17)xAI (8)Meta (6)DeepSeek (9)Mistral AI (11)Alibaba (Qwen) (34)Z.ai (12)Moonshot AI (5)
Popular modelsClaude Fable 5.1GPT-6 AstraClaude Opus 5Claude Fable 5GPT-5.6 SolGLM 5.3Grok 4.6Kimi K3GPT-5.6 Terra
By deliveryBatch APIPrompt cachingPriority tierFrontier tierProprietary APITool callingStructured JSONWeb searchMultilingual

Ratings 0

Avg. score 0

Rate this page

Help us improve