Identity
- Status
- active
- Developer ID
qwen3.8-max- Released
- Not reported
- Version
- Qwen3.8-Max
- License
- Unknown
QwenQwen3.8-Max
Qwen3.8-Max is Alibaba's 2.4T-parameter multimodal flagship for coding, office, and long-horizon agent tasks.
- Context
- 1,000,000 tokens
- Inputs
- Text, Image, Video
- Outputs
- Text
- Released
- Not reported
Providers & Pricing
Estimate workload cost →| Provider | Input / 1M | Output / 1M | Cache read / 1M | Cache write / 1M | Observed | Evidence |
|---|---|---|---|---|---|---|
 Alibaba Cloud Model StudioAvailable | $12.00 | $36.00 | $1.50 | — | Aug 29, 2026 | Provider-reported |
 DeepinfraAvailable | $1.65 | $4.951 | $0.206 | — | Sep 2, 2026 | Provider-reported |
 OpenrouterAvailable | $2.00 | $6.00 | $0.25 | — | Sep 2, 2026 | Provider-reported |
Benchmarks
All benchmark results →| Qwen3.8-MaxQwen | LiveBench2026-06-25 | 81.88 | 21,637 | 1,270 | qwen3.8-maxaverage per case | Recomputedwinner eligible | Third-party benchmark |
| Qwen3.8-MaxQwen | LMArena Agent Arenaagent-2026-08-31-011508720696 | 6.01 | — | 18,386 | Qwen3.8 Max; 95% CI [4.88117795, 7.13339178]; sessions 18386; observations 1372080; rank 14unknown | Original sourcewinner eligible | Third-party benchmark |
| Qwen3.8-MaxQwen | LMArena Text Arenatext-2026-09-01-011508720696 | 1,479.56 | — | 13,136 | qwen3.8-max; 95% CI [1473.34740586, 1485.77213629]; votes 13136; rank 12unknown | Original sourcewinner eligible | Third-party benchmark |
| Qwen3.8-MaxQwen | LMArena Vision Arenavision-2026-08-27-011508720696 | 1,312.86 | — | 7,244 | qwen3.8-max; 95% CI [1304.57707433, 1321.13715544]; votes 7244; rank 6unknown | Original sourcewinner eligible | Third-party benchmark |
| Qwen3.8-MaxQwen | ToneBench2026-08-14-10-task | 83.35 | 20,929 | 50 | Qwen3.8 Maxaverage per case | Original sourcewinner eligible | Third-party benchmark |
At a Glance
Model Facts
Capacity
- Context
- 1,000,000 tokens
- Maximum output
- 131,072 tokens
- Total parameters
- 2,400,000,000,000
- Active parameters
- Unknown
- Knowledge cutoff
- Not reported
Interface and Access
- Inputs
- Text, Image, Video
- Outputs
- Text
- API available
- Yes
- Open weights
- No
- Self-hostable
- No
- Reasoning
- Yes
- Vision
- Yes
- Tool calling
- Yes
- Capabilities
- agents, chat, reasoning, structured outputs, tools, vision
This page represents one immutable developer model ID. Serving endpoints, pricing, regions, quantizations, and service tiers attach separately and never create duplicate model records.
Version Lineage
PredecessorUnknown
SuccessorsUnknown
Aliasesqwen3.8-max
Explicit Unknowns
- Release date
- Tokenizer
- Knowledge cutoff
Primary Evidence
Sources and Observation Date
Comparable Peers
Related Model Comparisons
| A | Pair | B | Context |
|---|---|---|---|
Qwen3.8-MaxQwen | vs | Qwen3.7-PlusQwen | family variantsimage, text, video |
Qwen3.8-MaxQwen | vs | Qwen3.8-FlashQwen | family variantsimage, text, video |
Qwen3.8-MaxQwen | vs |  GLM-5V-TurboZ.ai | cross-developer peersimage, text, video |
Qwen3.8-MaxQwen | vs |  DeepSeek-V4-Flash-Vision-ExpDeepSeek | cross-developer peersimage, text |
Qwen3.8-MaxQwen | vs |  Kimi-K3Moonshot AI | cross-developer peersimage, text |
Qwen3.8-MaxQwen | vs |  Gemini 2.5 FlashGoogle DeepMind | cross-developer peerstext |
Qwen3.8-MaxQwen | vs | Gemini 3.6 FlashGoogle DeepMind | cross-developer peerstext |
Qwen3.8-MaxQwen | vs | Gemini 3.7 FlashGoogle DeepMind | cross-developer peerstext |
Qwen3.8-MaxQwen | vs | family variantstext | |
Qwen3.8-MaxQwen | vs |  Sonar Deep ResearchPerplexity | cross-developer peerstext |
Qwen3.8-MaxQwen | vs | Qwen3-Coder-NextQwen | family variantstext |
Qwen3.8-MaxQwen | vs | Gemini Deep Research MaxGoogle DeepMind | cross-developer peerstext |
Questions
Qwen3.8-Max FAQs
What is Qwen3.8-Max?+
Qwen3.8-Max is Alibaba's 2.4T-parameter multimodal flagship for coding, office, and long-horizon agent tasks. It is developed by Qwen and its current sourced lifecycle status is active.
What inputs and outputs does Qwen3.8-Max support?+
Qwen3.8-Max's current record lists Text, Image, and Video as input and Text as output.
How much context does Qwen3.8-Max support?+
The current record lists a 1,000,000-token context window and a maximum output of 131,072 tokens.
Is Qwen3.8-Max available through an API?+
Qwen3.8-Max is marked as API-available. The directly sourced provider records currently include Alibaba Cloud Model Studio, Deepinfra, and Openrouter.
How large is Qwen3.8-Max?+
The current record lists 2,400,000,000,000 total parameters and an unknown active parameter count. Its architecture is mixture_of_experts.
What capabilities and tasks does Qwen3.8-Max support?+
Qwen3.8-Max's recorded capabilities are agents, chat, reasoning, structured outputs, tools, and vision. Its supported tasks are coding, image, reasoning, text, and video.
How much does Qwen3.8-Max cost through an API?+
The lowest directly sourced prices currently attached to Qwen3.8-Max are $1.65 per million input tokens through Deepinfra and $4.951 per million output tokens through Deepinfra. Prices are provider-specific and should be checked against each cited observation date.
Can Qwen3.8-Max be self-hosted?+
Qwen3.8-Max is not marked as an open-weight model. Self-hosting is marked as unsupported, with no license recorded.
Does Qwen3.8-Max have related versions or aliases?+
No predecessor is recorded. No successor is recorded. Recorded aliases are qwen3.8-max.
Which models can Qwen3.8-Max be compared with?+
The current compatible peer set includes Qwen3.7-Plus, Qwen3.8-Flash, GLM-5V-Turbo, DeepSeek-V4-Flash-Vision-Exp, and Kimi-K3. Compatibility requires shared sourced modalities and tasks; it does not imply that one model is better. Browse model comparisons →