Qwen3.7 Max · Alibaba (Qwen)

What does Qwen3.7 Max cost to run in a coding agent?

Qwen3.7 Max costs about $0.063 per average coding-agent request on our measured traffic, at OpenRouter's listed prices of $1.475 per million input and $4.425 per million output tokens. It scores 29 on the Artificial Analysis Intelligence Index, 30th of the 48 models we track. Source: Artificial Analysis (artificialanalysis.ai)

List prices: OpenRouter listed price, checked 2026-10-13. Request costs: Standard Compute gateway data, 79,301 coding-agent requests from 75 users, Sep 13 to Oct 12, 2026. How we compute this

Run Qwen3.7 Max in your agent

Or let the router send each step to the cheapest model that can do it, on providers we vet for price, uptime and speed. Plans from $19 a month.

See plans

What Qwen3.7 Max costs per coding-agent request

Cost at list price	Qwen3.7 Max
Prompt under 100K tokens50% of requests, 87% of input cached	$0.029
Prompt of 100K to 272K39% of requests, 88% of input cached	$0.076
Prompt over 272K tokens11% of requests, 86% of input cached	$0.17
Average request	$0.063
100 requests	$6.29
Median session16 requests; a session ends after 10 idle minutes	$0.79
A month of regular usemedian of 20 users active on 5+ days	$103

Each logged request priced at the listed rates above for its own token counts, including cache reads and any higher rate for long prompts. Average output: 698 tokens per request.

Qwen3.7 Max benchmarks, speed and latency

Artificial Analysis	Value	Rank
Intelligence Index10 evaluations, 0 to 100	29	30 of 48
Terminal-Bench 4.0agentic coding in a terminal	1.5%	29 of 39
Output speedmedian	197 tokens/s	5 of 44
Time to first answer tokenincludes thinking time	14 s	6 of 44

Source: Artificial Analysis (artificialanalysis.ai). Measured variant: Qwen3.7 Max, speed on Alibaba. Fetched October 13, 2026. Rank is among the models on this site, older versions included. Artificial Analysis runs Terminal-Bench 4.0 in its own setup, so its scores differ from the tbench.ai leaderboard.

About Qwen3.7 Max

Qwen3.7 Max is Alibaba's hosted flagship Qwen3.7 model, released May 19, 2026 with a 1M-token context and no open weights. It is fast, at about 197 output tokens per second per Artificial Analysis, but it scores 30 on the Artificial Analysis Intelligence Index and 1.5% in its Terminal-Bench 4.0 run, well below current frontier models on agentic coding, and Artificial Analysis lists it as superseded. At OpenRouter's listed price its cached input costs 20% of the input rate, so in our measurements it costs more per request than Claude Sonnet 5.5 ($0.063 vs $0.053).

Sources: Artificial Analysis

Qwen3.7 Max FAQ

How much does Qwen3.7 Max cost per coding-agent request?

$0.063 on average at OpenRouter's listed price, $6.29 per 100 requests, and $0.79 for a median 16-request session. That is 79,301 real coding-agent requests (Sep 13 to Oct 12, 2026) priced at OpenRouter's listed rates; the median request sent 99,534 input tokens, 87% of input was read from cache, and requests returned 698 output tokens on average.

Does Qwen3.7 Max charge more for long prompts?

No. Qwen3.7 Max bills the same rate per token up to its 1M-token context window.

How fast is Qwen3.7 Max?

Artificial Analysis measures Qwen3.7 Max at 197 output tokens per second, with 14 seconds to the first answer token, thinking time included, on Alibaba's API. Source: Artificial Analysis (artificialanalysis.ai).

Is Qwen3.7 Max good for AI agents?

Fast, long-context general work in Qwen-based stacks. On our scoring, which is calibrated to public benchmarks, it rates 4.9/10 for agentic ability and 6.3/10 for output quality. 1M context.

What is Qwen3.7 Max's context window?

Qwen3.7 Max has a 1M-token context window, large enough to hold sizeable codebases or document sets in a single run. Alibaba (Qwen) is the maker.

Can I self-host Qwen3.7 Max?

No. Qwen3.7 Max is a proprietary, hosted-only model from Alibaba (Qwen). You access it through an API, such as OpenRouter, rather than running the weights yourself.

What are the downsides of Qwen3.7 Max?

The main trade-offs: Weak on long agentic tasks (1.5% in Artificial Analysis's Terminal-Bench 4.0 run); Costs more per request than Claude Sonnet 5.5 at OpenRouter's listed price; Hosted only, no open weights. It's strongest for fast, long-context general work in Qwen-based stacks.