Qwen3.5-9B
Qwen · released Feb 27, 2026 · Qwen/Qwen3.5-9B
- Type
- Open weightsApache License 2.0
- Params
- 9.7B
- Context
- 262K
about 197K words of context
Our take
Written Aug 4, 2026Alibaba's 9-billion-parameter small downloadable model has a permissive Apache licence and an unusually large request limit for its size. It is designed for consumer GPUs and cheap hosted inference.
Use this for local inference on 8–12GB video-memory GPUs, or cheap hosted multimodal input. Pick it for long-context small-model workloads. Skip it if you need measured quality scores or complex reasoning beyond the reach of a 9.7-billion-parameter model.
The case for it
- 9.7 billion parameters: a standard compressed size fits in 8GB-class video memory for local inference.
- Text, image and video input in a small model.
- 262,144-token request limit matches models many times larger.
The case against it
- No benchmark scores yet, so there is no measured quality data.
- Expect weaker complex reasoning than 24-billion-parameter peers.
How good is it?
We hold no score for this model.
So there is no figure here for everyday use, coding, agent work or writing. That is a gap in our data, not a low score.
Can you run it yourself?
Comfortable fit
GeForce RTX 4090 · 24 GB
Room to spare. 15 GB spare means a 10% error in the size would not change the answer.
GeForce RTX 5090 · 32 GB
Room to spare. 23 GB spare means a 10% error in the size would not change the answer.
Apple M1 (8-core GPU) · 16 GB
Room to spare. 4.2 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked between 2 hours and 8 hours ago — each listing carries its own date.
Cheapest of the 2 listings we can compare like for like — at 262K of context, out of 7 in the table below. One cheaper row there is outside that comparison: a different quantisation.
- per 1M tokens
- $0.10 in / $0.15 out
- Context served
- 262K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| Darkbloomfp4Through OpenRouter | $0.080 / $0.13checked 8 hours ago | 262K66K max reply | 19 tok/s | No | Yesunknown period | Unknown |
| DeepInfrabf16Direct and through OpenRouter | $0.10 / $0.15checked 2 hours ago | 262K82K max reply through OpenRouter | 11 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
| OpenRouterOpenRouter's own listing | $0.10 / $0.15checked 2 hours ago | 262K | not measured | Unknown | Unknown | Unknown |
| Venice AIfp8Through OpenRouter | $0.10 / $0.15checked 2 hours ago | 256K33K max reply | 45 tok/s | No | No | Confirmed |
| SiliconFlowfp8Through OpenRouter | $0.10 / $0.15checked 2 hours ago | 262K236K max reply | 19 tok/s | No | No | Confirmed |
| Parasailbf16Through OpenRouter | $0.10 / $0.25checked 2 hours ago | 262K236K max reply | 111 tok/s | No | No | Confirmed |
| Together AIThrough OpenRouter | $0.17 / $0.25checked 8 hours ago | 262K236K max reply | 81 tok/s | No | No | Confirmed |
DeepInfra
Together AIAcross the 7 listings we hold: 6 say they do not train on prompts (1 of them only through OpenRouter), 0 say they do and 1 does not say. 5 appear in the zero-retention registry we check (1 of them only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| Darkbloomfp4Through OpenRouter | ✓ | ✓ | ✓ |
| DeepInfrabf16Direct and through OpenRouter | ✓ | ✓ | ✓ |
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| Venice AIfp8Through OpenRouter | ✓ | ✓ | ✓ |
| SiliconFlowfp8Through OpenRouter | ✓ | ✓ | ✓ |
| Parasailbf16Through OpenRouter | ✗ | ✓ | ✓ |
| Together AIThrough OpenRouter | ✓ | ✓ | ✓ |
Tool calling: 6 of 7 listings say yes, 1 says no. JSON output: 7 of 7 listings say yes. Strict schema: 7 of 7 listings say yes.
Models people weigh against Qwen3.5-9B
When we formed this view
Recent changes
What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
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.
- No independent board has scored it, so we hold no quality figures at all.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 7 listings does not say whether it trains on prompts, and 1 answers only through OpenRouter, not for its own listing.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
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/Qwen3.5-9B
- Architecture
- Dense
- Takes in, gives back
- Text, images and video in, text out
- Catalogue slug
- qwen-qwen3-5-9b