Qwen3.7 Max
Qwen · released May 21, 2026
- Type
- Proprietary
- Input
- None held
- Output
- None held
- Cached
- None held
We don't hold a list price for this model yet · hosted only — we hold no downloadable copy
Our take
Written Aug 2, 2026Qwen 3.7 Max is Alibaba's proprietary flagship text model, released in May 2026 with a one-million-token request limit and strong measured coding performance. It is a premium long-context offering available through five providers, with a wide price spread between the cheapest and most expensive hosts.
Choose this for coding-heavy tasks where its measured Arena Coding score is strongest, or for long-document workflows needing one million tokens in a single prompt at below-frontier cost. Use it if you can route through the cheapest host, where the same model costs half what the most expensive charges. Skip it if you need downloadable weights, multimodal input, or reliable agentic behaviour.
The case for it
- Strong measured coding performance: its Arena Coding score sits about 50 points above its own overall text score.
- Wide price spread for identical access: the cheapest host charges half the input and output cost of the most expensive.
- One-million-token request limit available at every tier, including the cheapest offers.
- Alibaba's own listings show measured throughput of 46–48 tokens per second.
The case against it
- Agentic capability score is anomalously negative and likely erroneous, far below every other measured skill.
- Proprietary weights with no open licence; cannot be self-hosted or fine-tuned freely.
- Throughput undisclosed on three of the five tracked offers.
How good is it?
IntelligencePuzzles, maths, exam questions
Arena Text (overall)13th of 143 · 1474.7
CodingWriting and fixing code on its own
Arena Coding10th of 143 · 1525.3
AgenticPlanning, calling tools, staying on task
Arena Agent (IPS)20th of 36 · −0.005
WritingWe do not rate this
Two boards come close and neither tests writing: Arena Creative Writing asks people which of two replies they prefer, and LiveBench Language tests whether a model understood a passage. So we show where Qwen3.7 Max placed and give it no mark out of five.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done, which is why they get no rating.
Every published score for this model16 scoresEvery figure we hold, from 16 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Or rent it from someone else
Cheapest of 5 live listings. Picked at the widest standard context we hold, within one quantisation slice, so the numbers beside it are a price one host actually charges.
- per 1M tokens
- $1.25 in / $3.75 out
- Context served
- 1M
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| Novita AI | $1.25 / $3.75 | 1M | not measured | Unknown | Unknown | Unknown |
| OpenRouter | $1.48 / $4.42 | 1M | not measured | Unknown | Unknown | Unknown |
| Alibaba Cloudfp8 | $1.48 / $4.42 | 1M | 44 tok/s | No | Yesunknown period | Unknown |
| Alibaba Cloud | $1.48 / $4.42 | 1M | 46 tok/s | No | Yesunknown period | Unknown |
| DeepInfra | $2.50 / $7.50 | 256K | not measured | Unknown | Unknown | Unknown |
OpenRouter
DeepInfraAcross the 5 listings we hold: 2 say they do not train on prompts, 0 say they do and 3 do not say. 0 appear in the zero-retention registry we check; the rest are unknown to us rather than confirmed either way.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
- ✓
- Supported
- ✗
- Not supported
- Not published
- host gave no parameter list
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| Novita AI | |||
| OpenRouter | ✓ | ✓ | ✓ |
| Alibaba Cloudfp8 | ✓ | ✓ | ✓ |
| Alibaba Cloud | ✓ | ✓ | ✓ |
| DeepInfra |
Tool calling: 3 of 5 listings say yes, 2 publish no parameter list. JSON output: 3 of 5 listings say yes, 2 publish no parameter list. Strict schema: 3 of 5 listings say yes, 2 publish no parameter list.
Models people weigh against Qwen3.7 Max
When we formed this view
Dates behind this page
Prices last checked 9d ago
What we do not know about this model yet
- 2 of 5 listings publish no parameter list, so what their API accepts is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 3 of 5 listings do not say whether they train on prompts.
- We don't hold a list price for this model yet — the gap is ours, not the lab's.
Licence and identifiers
What the licence allowsWe hold no licence record for this model. Inside are the identifiers you need to pull it — its Hugging Face repo where we have one, our slug and a machine-readable card.
Licence
Commercial API terms. We hold no licence record for this model, so there is nothing to summarise here.
Identifiers
- Modality record
- text->text
- Catalogue slug
- qwen-qwen3-7-max