Qwen4-Max
Flagshipby Alibaba
As of November 16, 2026, Qwen4-Max by Alibaba costs $1.50 per 1M input tokens and $4.50 per 1M output tokens, with a 1M-token context window.
Qwen4-Max is the first generally purchasable model in Alibaba's Qwen 4 generation, and the way it arrived is as notable as the model. Alibaba opened Qwen4-Max-Preview on November 11, 2026 behind Token Plan subscriptions only, with no per-token price, no benchmarks, no parameter count, and a documented preview window closing on November 25. The pay-as-you-go model then appeared on Alibaba Cloud Model Studio without a launch post: the model pricing page now carries a qwen4-max row at $1.50 per million input tokens and $4.50 output in Singapore, the region basis the rest of the Qwen catalog here uses, and $1.413 and $4.24 in Beijing and the global regions. That undercuts Qwen3.8-Max at $2 and $6 by a quarter on both sides, which makes this the rare generational step that lowers the flagship rate rather than raising it. Alibaba's model page documents the shape in detail: a 1,000,000 token context length, up to 991,808 input tokens without thinking and 983,616 with it, 131,072 output tokens in non-thinking mode and up to 262,144 tokens combined across thinking and the final answer, text, image, and video input with text output, and thinking on by default at high effort with low, high, and xhigh available (the older minimal, medium, and max values map onto those three). It speaks the OpenAI Chat Completions and Responses protocols, the Anthropic Messages protocol, and DashScope, supports explicit and implicit caching, parallel tool calling, web search, structured output, and thinking summaries in every region, and supports neither fine-tuning nor model deployment anywhere. Batch inference is Beijing only. What is still missing is evidence: Alibaba has published no benchmark table, no parameter count, no knowledge cutoff, and no weights for any Qwen 4 model, and no independent evaluator has scored it yet, so treat the price and the spec sheet as the only verified facts here.
Input Price
$1.50
per 1M tokens
Output Price
$4.50
per 1M tokens
Context Window
1M
tokens
Released
2026-11
API access
Capabilities
Key Strengths
- ✓$1.50 / $4.50 per 1M tokens, a quarter under Qwen3.8-Max
- ✓1,000,000 token context length with up to 262,144 tokens of thinking plus answer
- ✓Text, image, and video input
- ✓Explicit and implicit prompt caching in every region
- ✓Speaks OpenAI Chat Completions and Responses, Anthropic Messages, and DashScope
- ✓Thinking on by default with low, high, and xhigh effort
Best For
- ▸Long-context agentic work at below Qwen3.8-Max cost
- ▸Multimodal document, chart, and video reasoning
- ▸Drop-in testing against an OpenAI or Anthropic-shaped client
- ▸Complex reasoning and coding where a Chinese-hosted flagship is acceptable
Pricing Details
Input tokens
$1.50
per 1M tokens
Output tokens
$4.50
per 1M tokens
Estimated cost per 1K requests
$3.75
~1K input + ~500 output tokens avg
Prices are subject to change. Check the official documentation for current pricing. See the cost calculator for detailed estimates.
Quick answers
- How much does Qwen4-Max cost?
- As of November 16, 2026, Qwen4-Max costs $1.50 per 1M input tokens and $4.50 per 1M output tokens on the Alibaba API.
- What is the context window of Qwen4-Max?
- Qwen4-Max has a 1M-token context window (1,000,000 tokens).
Related Models
Qwen3.8 27B
Mid-tierAlibaba
$0.50 in / $3.00 out
Qwen3.8-Omni-Flash
BudgetAlibaba
$0.15 in / $0.47 out
Qwen3.8-Flash
BudgetAlibaba
$0.15 in / $0.47 out
Qwen3.8 2.4T-A95B
FlagshipAlibaba
$2.00 in / $6.00 out
Qwen3.8-Max
FlagshipAlibaba
$2.00 in / $6.00 out
Qwen3.8 Max Prime
FlagshipAlibaba
$4.00 in / $12.00 out
Qwen3.7-Max
FlagshipAlibaba
$2.50 in / $7.50 out
Claude Fable 5.2
FlagshipAnthropic
$8.00 in / $40.00 out
Claude Mythos 5.2
FlagshipAnthropic
$8.00 in / $40.00 out
Claude Mythos 5.1
FlagshipAnthropic
$10.00 in / $50.00 out
Claude Opus 5.5
FlagshipAnthropic
$4.00 in / $20.00 out