Qwen3-VL Models by Alibaba
Last refreshed 2026-06-04. Next refresh: weekly.
Details
Capabilities
About
Qwen3-VL is Alibaba’s Qwen Team vision–language series for interleaved text, image, and video understanding. The family ships dense sizes (2B/4B/8B/32B) and MoE sizes (30B-A3B and flagship 235B-A22B), each with Instruct and Thinking post-trained editions and a native long multimodal context window. Qwen documents stronger visual agents (GUI operation), spatial grounding, OCR, STEM reasoning, and long video/document work, with architecture updates such as Interleaved-MRoPE, DeepStack vision fusion, and text-timestamp video alignment.
Decision facts
- Best fit
- vision and multimodal workJSON / Tool usestructured outputs
- Capability starting point
- Qwen3-VL-235B-A22B with 128k context and structured outputs and multimodal inputs
- Lowest tracked input
- Qwen3 VL 8B Instruct · $0.080/1M · Novita AI
- Closest related family
- Ovis Image
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs 128k context, structured outputs, and multimodal inputs.
Use when the workload needs 256k context, 235B parameters, and JSON / Tool use.
Use when the workload needs 128k context, 30B parameters, and JSON / Tool use.
Use when the workload needs 128k context, 32B parameters, and JSON / Tool use.
Use when the workload needs 128k context, 8B parameters, and JSON / Tool use.
Use when the workload needs multimodal, 128k context, and 30B parameters.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Qwen3-VL-235B-A22B | Use when the workload needs 128k context, structured outputs, and multimodal inputs. | 2025-12 | 128k contextstructured outputsmultimodal inputs | Current |
| Qwen3 VL 235B A22B Instruct | Use when the workload needs 256k context, 235B parameters, and JSON / Tool use. | 2025-09 | 256k context235B parametersJSON / Tool use | Current |
| Qwen3 VL 30B A3B Instruct | Use when the workload needs 128k context, 30B parameters, and JSON / Tool use. | 2025-09 | 128k context30B parametersJSON / Tool use | Current |
| Qwen3 VL 32B Instruct | Use when the workload needs 128k context, 32B parameters, and JSON / Tool use. | 2025-09 | 128k context32B parametersJSON / Tool use | Current |
| Qwen3 VL 8B Instruct | Use when the workload needs 128k context, 8B parameters, and JSON / Tool use. | 2025-09 | 128k context8B parametersJSON / Tool use | Current |
| Qwen3-VL-30B-A3B | Use when the workload needs multimodal, 128k context, and 30B parameters. | 2025-01 | multimodal128k context30B parameters | Current |
Release Timeline
3 release groupsSpecifications(6 models)
| Model | Released | Context | Parameters | Vision | Multimodal | JSON / Tool use | Structured Outputs |
|---|---|---|---|---|---|---|---|
| Qwen3-VL-235B-A22B | 2025-12 | 128k | 235B (22B active) | Yes | Yes | No | Yes |
| Qwen3 VL 235B A22B Instruct | 2025-09 | 256k | 235B | Yes | Yes | Yes | Yes |
| Qwen3 VL 30B A3B Instruct | 2025-09 | 128k | 30B | Yes | Yes | Yes | Yes |
| Qwen3 VL 32B Instruct | 2025-09 | 128k | 32B | Yes | Yes | Yes | Yes |
| Qwen3 VL 8B Instruct | 2025-09 | 128k | 8B | Yes | Yes | Yes | Yes |
| Qwen3-VL-30B-A3B | 2025-01 | 128k | 30B | Yes | Yes | No | Yes |
Available From(5 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Qwen3 VL 8B Instruct | Novita AI | $0.08 | $0.5 | Serverless |
| Qwen3 VL 32B Instruct | OpenRouter | $0.104 | $0.416 | Serverless |
| Qwen3 VL 8B Instruct | OpenRouter | $0.117 | $0.455 | Serverless |
| Qwen3-VL-30B-A3B | OpenRouter | $0.13 | $1.56 | Serverless |
| Qwen3 VL 30B A3B Instruct | OpenRouter | $0.13 | $0.52 | Serverless |
| Qwen3-VL-30B-A3B | Fireworks AI | $0.15 | $0.6 | Serverless |
| Qwen3 VL 30B A3B Instruct | Novita AI | $0.2 | $0.7 | Serverless |
| Qwen3 VL 235B A22B Instruct | OpenRouter | $0.21 | $1.9 | Serverless |
| Qwen3-VL-235B-A22B | OpenRouter | $0.26 | $2.6 | Serverless |
| Qwen3 VL 235B A22B Instruct | Novita AI | $0.3 | $1.5 | Serverless |
| Qwen3 VL 235B A22B Instruct | Vercel AI Gateway | $0.4 | $1.6 | Serverless |
| Qwen3-VL-235B-A22B | AWS Bedrock | $0.53 | $2.66 | Serverless |
| Qwen3-VL-235B-A22B | Novita AI | $0.98 | $3.95 | Serverless |
Related Model Families
Models(6)
Qwen3-VL-235B-A22B
Qwen3 VL 235B A22B Instruct
Qwen3 VL 30B A3B Instruct
Qwen3 VL 32B Instruct
Qwen3 VL 8B Instruct
Qwen3-VL-30B-A3B