Qwen3.5 397B A17B
Qwen · released Feb 16, 2026 · Qwen/Qwen3.5-397B-A17B
- Type
- Open weightsApache License 2.0
- Params
- 403B
- Context
- 262K
17B active per word · about 197K words of context
Our take
Written Aug 3, 2026Qwen 3.5 is a large downloadable model from Alibaba with a permissive Apache licence and a mixture-of-experts design that keeps 17 billion parameters active per token. It scores near the top on measured coding and hard-prompt reasoning among open-weights models, and accepts text, images and video.
Choose this for frontier-level coding with a permissive licence, or for hard-prompt reasoning tasks where measured scores matter. It suits cost-sensitive multimodal inference and teams who want to self-host or fine-tune a large model. Skip it if creative writing or web-development coding is your main workload, or if you need guaranteed high throughput and cannot shop between providers.
The case for it
- Top-tier measured coding ability among downloadable models: Arena Coding Elo 1490.8.
- Strong on deliberately difficult prompts: Arena Hard Prompts Elo 1463.3.
- Permissive Apache 2.0 licence with 403.4 billion total parameters and 17 billion active per token.
- Wide provider choice with competitive rates and throughput ranging from 4 to 63 tokens per second.
The case against it
- Creative writing lags 84 points behind its own coding score on the same leaderboard.
- Web-development coding is 91 points below its general coding peak.
- Throughput varies sixteenfold across providers for the same model, so performance depends heavily on host choice.
How good is it?
IntelligencePuzzles, maths, exam questions
Arena Text (overall)44th of 143 · 1441.7
CodingWriting and fixing code on its own
Arena Coding44th of 143 · 1490.8
AgenticPlanning, calling tools, staying on task
Nobody we watch has scored Qwen3.5 397B A17B for this. We would take the rating from Arena Agent (IPS).
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 Qwen3.5 397B A17B 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 model7 scoresEvery figure we hold, from 7 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
- Fits in memory
- weights load entirely on the card
- Spills to system RAM
- some weights offload; much slower
- Too large
- will not load even with offload
- est
- size is calculated; the verdict could change by 10%
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Comfortable fit
Apple M3 Ultra (80-core GPU) · 512 GB
Room to spare. 120.6 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
All 3 sizes here are calculated, not measured. We hold no measured file for this model, so each size comes from the parameter count and every verdict above inherits that uncertainty.
Check against your own machine → · All 71 devices, with every size →
Or rent it from someone else
Cheapest of 15 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
- $0.39 in / $2.34 out
- Context served
- 262K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouter | $0.39 / $2.34 | 262K | not measured | Unknown | Unknown | Unknown |
| Alibaba Cloud | $0.39 / $2.34 | 262K | 63 tok/s | No | Yesunknown period | Unknown |
| Alibaba Cloudfp8 | $0.39 / $2.34 | 262K | 34 tok/s | No | Yesunknown period | Unknown |
| DigitalOcean Gradient | $0.39 / $2.45 | 131K | 4 tok/s | No | No | Confirmed |
| DeepInfrafp8 | $0.45 / $3.00 | 262K | 27 tok/s | No | No | Confirmed |
| DeepInfrafp8 | $0.45 / $3.00 | 262K | not measured | Unknown | Unknown | Unknown |
| Chutesfp8 | $0.45 / $3.00 | 262K | 5 tok/s | No | Yesunknown period | Unknown |
| AtlasCloudfp8 | $0.55 / $3.50 | 262K | 35 tok/s | No | Yesunknown period | Unknown |
| Phala | $0.55 / $3.50 | 262K | 48 tok/s | No | No | Confirmed |
| Novita AI | $0.60 / $3.60 | 262K | not measured | Unknown | Unknown | Unknown |
| Novita AI | $0.60 / $3.60 | 262K | 47 tok/s | No | No | Confirmed |
| Parasailfp8 | $0.50 / $3.60 | 262K | 49 tok/s | No | No | Confirmed |
| GMICloudfp8 | $0.60 / $3.60 | 262K | not measured | No | Yesunknown period | Unknown |
| StreamLake | $0.60 / $3.60 | 256K | 9 tok/s | No | Yesunknown period | Unknown |
| Venice AI | $0.75 / $4.50 | 128K | 49 tok/s | No | No | Confirmed |
DigitalOcean Gradient
DeepInfra
Novita AIAcross the 15 listings we hold: 12 say they do not train on prompts, 0 say they do and 3 do not say. 6 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 |
|---|---|---|---|
| OpenRouter | ✓ | ✓ | ✓ |
| Alibaba Cloud | ✓ | ✓ | ✗ |
| Alibaba Cloudfp8 | ✓ | ✓ | ✗ |
| DigitalOcean Gradient | ✓ | ✓ | ✓ |
| DeepInfrafp8 | ✓ | ✓ | ✓ |
| DeepInfrafp8 | |||
| Chutesfp8 | ✓ | ✓ | ✓ |
| AtlasCloudfp8 | ✓ | ✓ | ✓ |
| Phala | ✓ | ✗ | ✓ |
| Novita AI | |||
| Novita AI | ✓ | ✓ | ✗ |
| Parasailfp8 | ✓ | ✓ | ✓ |
| GMICloudfp8 | ✓ | ✓ | ✗ |
| StreamLake | ✗ | ✓ | ✓ |
| Venice AI | ✓ | ✓ | ✓ |
Tool calling: 12 of 15 listings say yes, 1 says no, 2 publish no parameter list. JSON output: 12 of 15 listings say yes, 1 says no, 2 publish no parameter list. Strict schema: 9 of 15 listings say yes, 4 say no, 2 publish no parameter list.
Models people weigh against Qwen3.5 397B A17B
When we formed this view
Dates behind this page
Prices last checked 37h ago
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- 2 of 15 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 15 listings do not say whether they train on prompts.
Licence and identifiers
What the licence allowsApache License 2.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Apache License 2.0
Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.
Identifiers
- Hugging Face
- Qwen/Qwen3.5-397B-A17B
- Architecture
- Mixture of experts
- Modality record
- text+image+video->text
- Catalogue slug
- qwen-qwen3-5-397b-a17b