Alibaba (Qwen)

Qwen3.5 27B

Generally availableOpen weights262K contextReasoning

Summary

Qwen3.5 27B 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 $0.25 per million input tokens and $2 per million output tokens (SiliconFlow), across 6 providers we track.

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]
Parameters27.8B [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]
Hugging FaceQwen/Qwen3.5-27B [Hugging Face]

Qwen3.5 27B API pricing by provider

USD per million tokens. Sorted by input price.

ProviderInputOutputCached inputContextSource
SiliconFlow
Qwen/Qwen3.5-27B
$0.25$2–262Kmodels.dev, 18 Nov 2026
DeepInfra
Qwen/Qwen3.5-27B
$0.26$2.60–262Kmodels.dev, 18 Nov 2026
OpenRouter
qwen/qwen3.5-27b
$0.26$2.60–262KOpenRouter, 18 Nov 2026
Alibaba Cloud Model Studio
qwen3.5-27b
$0.30$2.40–262Kmodels.dev, 18 Nov 2026
Hugging Face Inference Providers
Qwen/Qwen3.5-27B
$0.30$2.40–262Kmodels.dev, 18 Nov 2026
Novita AI
qwen/qwen3.5-27b
$0.30$2.40–262Kmodels.dev, 18 Nov 2026

Running Qwen3.5 27B 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)17 GB18 GB20 GB
Q8_0 (8-bit)30 GB31 GB32 GB
FP16 / BF16 (16-bit)56 GB57 GB59 GB

At 4-bit with an 8K conversation, Qwen3.5 27B needs about 18 GB, so it fits on a graphics card with 24 GB or more, such as the NVIDIA GeForce RTX 3090 24 GB. Check your own hardware, or see Best GPU for local LLMs on our sister site AIPCs.

Frequently asked questions

What is the context window of Qwen3.5 27B?

Qwen3.5 27B 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 Qwen3.5 27B API cost?

Through Alibaba Cloud Model Studio, Qwen3.5 27B costs $0.30 per million input tokens and $2.40 per million output tokens. The lowest listed price is $0.25 input / $2 output through SiliconFlow. Prices last verified 18 Nov 2026.

Is Qwen3.5 27B open source?

Qwen3.5 27B is an open-weight model: its weights are published on Hugging Face as Qwen/Qwen3.5-27B under the apache-2.0 license. Check the license terms before commercial use.

When was Qwen3.5 27B released?

Alibaba (Qwen) released Qwen3.5 27B on 23 Feb 2026, according to models.dev.

Which providers offer Qwen3.5 27B?

We track Qwen3.5 27B on 6 providers: SiliconFlow, DeepInfra, OpenRouter, Alibaba Cloud Model Studio, Hugging Face Inference Providers, Novita AI.

How much VRAM does Qwen3.5 27B need?

About 18 GB at 4-bit (Q4_K_M) with an 8K-token conversation, and about 31 GB at 8-bit. The 4-bit model fits on a graphics card with 24 GB or more, such as the NVIDIA GeForce RTX 3090 24 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
Kimi K2.5Moonshot AI27 Jan 2026262K$0.45$2.25Open
GLM-4.7Zhipu AI (Z.ai)22 Dec 2025205K$0.40$1.75Open
GLM-5Zhipu AI (Z.ai)12 Feb 2026205K$0.95$2.55Open
Kimi K2.6Moonshot AI20 Apr 2026262K$0.65$3.40Open
Step 3.7 FlashStepFun29 May 2026256K$0.20$1.15Open
Macaron V1 TallMindlab21 Jul 2026262K$0.25$1.43Open

Change history

  • OpenRouter price changed: input $0.27 → $0.26, output $2.16 → $2.60 per 1M tokens
  • Input modalities changed from text, image, video, audio to text, image, video
  • OpenRouter price changed: input $0.26 → $0.27, output $2.60 → $2.16 per 1M tokens
  • OpenRouter price changed: input $0.20 → $0.26, output $1.56 → $2.60 per 1M tokens

Sources