Qwen3.7 Max
currentQwen3.7 Max is Alibaba's proprietary flagship for text-only reasoning, coding, office work, and long-horizon agents. It offers a one-million-token context and competitive instruction following at moderate prices, but it cannot inspect images and OpenRouter exposes reasoning primarily as an on/off choice rather than a fine-grained effort ladder.
Context window
1M tokens
Maximum output
66K tokens
Input price
$1.25 / 1M
Output price
$3.75 / 1M
Best for
- Text-only coding and agent workflows
- Long documents and repositories
- Instruction-heavy structured tasks
- Teams already using Alibaba or OpenRouter APIs
Not ideal for
- Image, audio, video, or PDF-vision inputs
- Applications needing multiple reasoning-effort levels
- Multi-provider failover on OpenRouter
- Use cases that require open weights
PromptHQ 100-word challenge
We give every model the same miniature science-fiction mystery. The response is short enough to read in one chat bubble, while the premise tests creative writing, compression, foreshadowing, causal coherence, and instruction following.
Qwen3.7 Max100 wordsWrite a science-fiction mystery in exactly 100 words.
A three-person research station on Europa receives a message from itself, timestamped 11 minutes in the future:
"Do not open the greenhouse door."
The door is opened anyway.
Plant two subtle clues. Give the warning a logical explanation without aliens, time travel, hallucinations, or simulations. The final sentence must change what the reader thought the warning meant.
Output only the story. Do not explain it.
Miller tapped the chronometer, still stuck eleven minutes fast after the power surge. On Europa, an automated console log blinked: Do not open the greenhouse door.
"System glitch," Sarah muttered. Her breath plumed in the uncharacteristically freezing corridor. She grabbed the manual override. "The crops need harvesting."
"Wait," Jax said, noticing the heavy magnetic deadbolts were engaged on the corridor side.
Sarah yanked the lever anyway.
The heavy steel door slid back, instantly sucking them into the silent vacuum of the shattered hydroponics bay, proving the automated system warning was never meant to protect the crew from the toxic plants.
Run the same prompt yourself
Open Qwen3.7 Max in Multichat with the evaluation prompt prefilled.
What the response revealed
The response contains exactly 100 whitespace-delimited words.
It outputs only the story and avoids all prohibited explanation types.
The fast chronometer explains the future timestamp, and the freezing corridor, external deadbolts, and breached bay explain the warning.
Visible breath and deadbolts on the corridor side are useful, understated clues to vacuum beyond the door.
The ending rejects the assumed plant hazard, though it states the reinterpretation more mechanically than dramatically.
The narrative is clear and complete, with minor repetition and a somewhat explanatory final line.
The model receives the prompt without web access or external tools. Qwen3.7 Max is run through OpenRouter with reasoning enabled at the provider default because this route exposes an on/off control rather than discrete effort levels. We preserve the response as generated apart from display rendering.
Performance
Qwen3.7 Max benchmarks
Benchmark scores are sensitive to reasoning effort, harness, tools, token budget, prompt format, sampling, and evaluation date. Scores here retain their source and should not be treated as directly interchangeable unless the underlying setup matches.
SWE-Bench Pro
60.6%
software engineering · Artificial Analysis3
SWE-Bench Verified
80.4%
software engineering · Artificial Analysis3
IFBench
79.1%
instruction following · Artificial Analysis3
GPQA Diamond
90.2%
graduate science reasoning · Artificial Analysis3
Terminal-Bench 2.1
74.5%
terminal and agentic coding · Artificial Analysis3
Family position
Qwen3.7 positioning
Max is the capability-focused Qwen3.7 model. It targets agents and long-context text work, while Plus and smaller Qwen variants cover lower-cost or multimodal use cases.
Qwen3.7 Max
This modelFlagship text agent
$1.25 input
$3.75 output
GLM 5.2
Open-weight coding rival
$0.93 input
$3 output
MiniMax M3
Lower-cost multimodal rival
$0.30 input
$1.20 output
API pricing
Per million text tokens
Input
$1.25
Cached input
$0.25
Cache write
$1.56
Output
$3.75
Where Qwen3.7 Max stands out
Built for demanding work
Text-only coding and agent workflows are central to the model's positioning, rather than an incidental capability.
Long-context capacity
The published context window is 1,000,000 tokens, making the model a candidate for large documents, repositories, and sustained agent state.
Reasoning and tools
Reasoning is supported with on, off provider setting(s), and the model can participate in tool-using workflows through its available API surface.
Limitations to know
Benchmarks are configuration-sensitive
Scores can move substantially with the harness, tool access, effort setting, token budget, and evaluator. Treat the table as evidence, not a universal ranking.
Context size is not guaranteed recall
A large advertised window does not mean every detail is retrieved reliably at maximum length. Validate representative long-context workloads before deployment.
Product access differs from model capability
PromptHQ and gateway limits may expose fewer modalities, tools, or tokens than the provider's first-party API.
Capabilities and specifications
Knowledge cutoff
Not publicly disclosed
Inputs
text
Reasoning
Supported
Default effort
provider-default
Supported API features
Frequently Asked Questions
Sources
- 1Qwen developer documentation
Alibaba Cloud · official documentation
- 2Qwen3.7 Max
OpenRouter · gateway model page
- 3Qwen3.7 Max intelligence, performance and price analysis
Artificial Analysis · independent benchmark
- 4PromptHQ model registry
PromptHQ · internal product configuration
Compare leading AI assistants
See how Qwen3.7 Max handles your own work.
Try Qwen3.7 Max in MultichatAvailable on PromptHQ Almostfree, Plus, Max