Alibaba (Qwen)

Qwen/Qwen3.5-9B

Generally availableOpen weights262K contextReasoning

Summary

Qwen/Qwen3.5-9B is an open-weight language model from Alibaba (Qwen), released on 23 Feb 2026. It accepts text, images and video and generates text, with a 262K-token context window and up to 66K output tokens per response. As of 18 Nov 2026, the lowest listed API price is $0.10 per million input tokens and $0.15 per million output tokens (DeepInfra), across 5 providers we track. It is scheduled for retirement on 21 Oct 2026.

Qwen vision-language model for visual reasoning, documents, and agent tasks

Last verified 18 Nov 2026. Every figure on this page links to its source.

Specifications

DeveloperAlibaba (Qwen)
Release date23 Feb 2026 [models.dev]
StatusGenerally available [models.dev]
WeightsOpen weights [Hugging Face]
Licenseapache-2.0 [Hugging Face]
Parameters9.7B [Hugging Face]
Context window262K (262,144 tokens) [models.dev]
Max output66K tokens [models.dev]
Knowledge cutoffApr 2025 [models.dev]
InputText, images and video [models.dev]
OutputText [models.dev]
Reasoning modeYes [models.dev]
Tool callingYes [models.dev]
Structured outputYes [models.dev]
Retirement date21 Oct 2026 [OpenRouter]
Hugging FaceQwen/Qwen3.5-9B [Hugging Face]

Qwen/Qwen3.5-9B API pricing by provider

USD per million tokens. Sorted by input price.

ProviderInputOutputCached inputContextSource
DeepInfra
Qwen/Qwen3.5-9B
$0.10$0.15–262Kmodels.dev, 18 Nov 2026
SiliconFlow
Qwen/Qwen3.5-9B
$0.10$0.15–262Kmodels.dev, 18 Nov 2026
Hugging Face Inference Providers
Qwen/Qwen3.5-9B
$0.17$0.25–262Kmodels.dev, 18 Nov 2026
Together AI
Qwen/Qwen3.5-9B
$0.17$0.25–262Kmodels.dev, 18 Nov 2026
OpenRouter
qwen/qwen3.5-9b
$0.17$0.25$0.15262KOpenRouter, 18 Nov 2026

Running Qwen/Qwen3.5-9B locally

Estimated memory, from the parameter count and the architecture in config.json on Hugging Face. Includes about 0.8 GB of runtime overhead.

QuantizationWeightsTotal, 8K contextTotal, 32K context
Q4_K_M (4-bit)5.9 GB6.9 GB7.7 GB
Q8_0 (8-bit)10 GB11 GB12 GB
FP16 / BF16 (16-bit)19 GB20 GB21 GB

At 4-bit with an 8K conversation, Qwen/Qwen3.5-9B needs about 6.9 GB, so it fits on a graphics card with 8 GB or more, such as the NVIDIA GeForce RTX 4060 8 GB. Check your own hardware, or see Running local LLMs on an 8 GB GPU on our sister site AIPCs.

Frequently asked questions

What is the context window of Qwen/Qwen3.5-9B?

Qwen/Qwen3.5-9B has a context window of 262,144 tokens (262K) and can generate up to 65,536 tokens in a single response, according to models.dev.

How much does the Qwen/Qwen3.5-9B API cost?

Qwen/Qwen3.5-9B is listed from $0.10 per million input tokens and $0.15 per million output tokens (DeepInfra). Prices last verified 18 Nov 2026.

Is Qwen/Qwen3.5-9B open source?

Qwen/Qwen3.5-9B is an open-weight model: its weights are published on Hugging Face as Qwen/Qwen3.5-9B under the apache-2.0 license. Check the license terms before commercial use.

When was Qwen/Qwen3.5-9B released?

Alibaba (Qwen) released Qwen/Qwen3.5-9B on 23 Feb 2026, according to models.dev.

Which providers offer Qwen/Qwen3.5-9B?

We track Qwen/Qwen3.5-9B on 5 providers: DeepInfra, SiliconFlow, Hugging Face Inference Providers, Together AI, OpenRouter.

How much VRAM does Qwen/Qwen3.5-9B need?

About 6.9 GB at 4-bit (Q4_K_M) with an 8K-token conversation, and about 11 GB at 8-bit. The 4-bit model fits on a graphics card with 8 GB or more, such as the NVIDIA GeForce RTX 4060 8 GB. These are estimates from the model's published parameter count and architecture.

Similar models from other labs

Comparable on output type, weights, release date, price and context window. Prices are the lowest listed, USD per million tokens.

ModelLabReleasedContextInput $/1MOutput $/1MWeights
NVIDIA Nemotron Nano 3 30BNVIDIA15 Dec 2025262K$0.05$0.20Open
Step 3.5 FlashStepFun29 Jan 2026262K$0.10$0.30Open
Gemma 4 26B A4B ITGoogle DeepMind2 Apr 2026262K$0.07$0.30Open
Ministral 3 8BMistral AI2 Dec 2025262K$0.15$0.15Open
NVIDIA Nemotron 3 Super 120B A12BNVIDIA11 Mar 2026262K$0.07$0.40Open
Ministral 14BMistral AI2 Dec 2025262K$0.20$0.20Open

Change history

  • OpenRouter price changed: input $0.08 → $0.17, output $0.12 → $0.25 per 1M tokens
  • Input modalities changed from text, image to text, image, video
  • Name changed from Qwen3.5 9B to Qwen/Qwen3.5-9B
  • Input modalities changed from text, image, video to text, image
  • Name changed from Qwen/Qwen3.5-9B to Qwen3.5 9B
  • Name changed from Qwen3.5 9B to Qwen/Qwen3.5-9B
  • Input modalities changed from text, image to text, image, video
  • OpenRouter price changed: input $0.17 → $0.08, output $0.25 → $0.12 per 1M tokens
  • Input modalities changed from text, image, video to text, image
  • OpenRouter price changed: input $0.04 → $0.17, output $0.07 → $0.25 per 1M tokens
  • OpenRouter price changed: input $0.17 → $0.04, output $0.25 → $0.07 per 1M tokens
  • OpenRouter price changed: input $0.17 → $0.17, output $0.25 → $0.25 per 1M tokens
  • Name changed from Qwen/Qwen3.5-9B to Qwen3.5 9B
  • Name changed from Qwen3.5 9B to Qwen/Qwen3.5-9B
  • Max output changed from 262K to 66K
  • Reasoning mode changed from no to yes
  • Input modalities changed from text, image to text, image, video
  • Name changed from Qwen/Qwen3.5-9B to Qwen3.5 9B
  • Input modalities changed from text, image, video to text, image
  • Reasoning mode changed from yes to no
  • Max output changed from 66K to 262K
  • Name changed from Qwen3.5 9B to Qwen/Qwen3.5-9B
  • Input modalities changed from text, image to text, image, video
  • Input modalities changed from text, image, video to text, image
  • Name changed from Qwen/Qwen3.5-9B to Qwen3.5 9B

Sources