Qwen3.5 Models by Alibaba
14 models2025–2026Up to 1m ctxFrom $0.07/1M input
Details
ResearcherAlibaba
LicenseApache 2.0OSI-approved
Commercial useCommercial use: permitted
Models14
Released2025–2026
Max context1m
Capabilities
Vision12 of 14 models
Multimodal12 of 14 models
Reasoning5 of 14 models
Function Calling5 of 14 models
Tool Use6 of 14 models
Structured Outputs5 of 14 models
About
Qwen3.5 is a family of 14 AI models by Alibaba, released between 2025 and 2026.
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
14 in view
Qwen-AgentWorld-35B-A3BCurrent
Use when the workload needs agents, 262k context, and reasoning.
2026-06agents262k contextreasoning
Qwen3.5-9BCurrent
Use when the workload needs 262k context, 9B parameters, and tool use.
2026-03262k context9B parameterstool use
Qwen3.5-4BCurrent
Use when the workload needs 262k context, 4B parameters, and multimodal inputs.
2026-03262k context4B parametersmultimodal inputs
Qwen3.5-2BCurrent
Use when the workload needs 262k context, 2B parameters, and multimodal inputs.
2026-03262k context2B parametersmultimodal inputs
Qwen3.5-0.8BCurrent
Use when the workload needs 262k context, 800M parameters, and multimodal inputs.
2026-03262k context800M parametersmultimodal inputs
Qwen3.5-27BCurrent
Use when the workload needs 262k context, 27B parameters, and reasoning.
2026-02262k context27B parametersreasoning
Qwen3.5-35B-A3BCurrent
Use when the workload needs 262k context, 35B parameters, and reasoning.
2026-02262k context35B parametersreasoning
Qwen3.5-122B-A10BCurrent
Use when the workload needs 262k context, 122B parameters, and reasoning.
2026-02262k context122B parametersreasoning
Qwen3.5-FlashCurrent
Use when the workload needs 1m context and multimodal inputs.
2026-021m contextmultimodal inputs
Qwen3.5-397B-A17BCurrent
Use when the workload needs 262k context, 397B parameters, and reasoning.
2026-02262k context397B parametersreasoning
Qwen3.5-PlusCurrent
Use when the workload needs 1m context and multimodal inputs.
2026-021m contextmultimodal inputs
Qwen3.5-4B-InstructCurrent
Use when the workload needs 256k context, 4B parameters, and multimodal inputs.
2025-11256k context4B parametersmultimodal inputs
Qwen3.5-32BCurrent
Use when the workload needs 256k context, 32B parameters, and multimodal inputs.
2025-11256k context32B parametersmultimodal inputs
Qwen3.5-32B-InstructCurrent
Use when the workload needs 256k context, 32B parameters, and multimodal inputs.
2025-11256k context32B parametersmultimodal inputs
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Qwen-AgentWorld-35B-A3B | Use when the workload needs agents, 262k context, and reasoning. | 2026-06 | agents262k contextreasoning | Current |
| Qwen3.5-9B | Use when the workload needs 262k context, 9B parameters, and tool use. | 2026-03 | 262k context9B parameterstool use | Current |
| Qwen3.5-4B | Use when the workload needs 262k context, 4B parameters, and multimodal inputs. | 2026-03 | 262k context4B parametersmultimodal inputs | Current |
| Qwen3.5-2B | Use when the workload needs 262k context, 2B parameters, and multimodal inputs. | 2026-03 | 262k context2B parametersmultimodal inputs | Current |
| Qwen3.5-0.8B | Use when the workload needs 262k context, 800M parameters, and multimodal inputs. | 2026-03 | 262k context800M parametersmultimodal inputs | Current |
| Qwen3.5-27B | Use when the workload needs 262k context, 27B parameters, and reasoning. | 2026-02 | 262k context27B parametersreasoning | Current |
| Qwen3.5-35B-A3B | Use when the workload needs 262k context, 35B parameters, and reasoning. | 2026-02 | 262k context35B parametersreasoning | Current |
| Qwen3.5-122B-A10B | Use when the workload needs 262k context, 122B parameters, and reasoning. | 2026-02 | 262k context122B parametersreasoning | Current |
| Qwen3.5-Flash | Use when the workload needs 1m context and multimodal inputs. | 2026-02 | 1m contextmultimodal inputs | Current |
| Qwen3.5-397B-A17B | Use when the workload needs 262k context, 397B parameters, and reasoning. | 2026-02 | 262k context397B parametersreasoning | Current |
| Qwen3.5-Plus | Use when the workload needs 1m context and multimodal inputs. | 2026-02 | 1m contextmultimodal inputs | Current |
| Qwen3.5-4B-Instruct | Use when the workload needs 256k context, 4B parameters, and multimodal inputs. | 2025-11 | 256k context4B parametersmultimodal inputs | Current |
| Qwen3.5-32B | Use when the workload needs 256k context, 32B parameters, and multimodal inputs. | 2025-11 | 256k context32B parametersmultimodal inputs | Current |
| Qwen3.5-32B-Instruct | Use when the workload needs 256k context, 32B parameters, and multimodal inputs. | 2025-11 | 256k context32B parametersmultimodal inputs | Current |
Release Timeline
4 release groups2026-06
1 current
Qwen-AgentWorld-35B-A3B
Currentagents262k contextreasoning
2026-03
4 current
Qwen3.5-0.8B
Current262k context800M parametersmultimodal inputs
Qwen3.5-2B
Current262k context2B parametersmultimodal inputs
Qwen3.5-4B
Current262k context4B parametersmultimodal inputs
Qwen3.5-9B
Current262k context9B parameterstool use
2026-02
6 current
Qwen3.5-122B-A10B
Current262k context122B parametersreasoning
Qwen3.5-27B
Current262k context27B parametersreasoning
Qwen3.5-35B-A3B
Current262k context35B parametersreasoning
Qwen3.5-397B-A17B
Current262k context397B parametersreasoning
Qwen3.5-Flash
Current1m contextmultimodal inputs
Qwen3.5-Plus
Current1m contextmultimodal inputs
2025-11
3 current
Qwen3.5-32B
Current256k context32B parametersmultimodal inputs
Qwen3.5-32B-Instruct
Current256k context32B parametersmultimodal inputs
Qwen3.5-4B-Instruct
Current256k context4B parametersmultimodal inputs
Specifications(14 models)
| Model | Released | Context | Parameters | Vision | Multimodal | Reasoning | Fn Calling | Tool Use | Structured Outputs |
|---|---|---|---|---|---|---|---|---|---|
| Qwen-AgentWorld-35B-A3B | 2026-06 | 262k | 35B total, 3B active | No | No | Yes | No | Yes | No |
| Qwen3.5-9B | 2026-03 | 262k | 9B | Yes | Yes | No | Yes | Yes | Yes |
| Qwen3.5-4B | 2026-03 | 262k | 4B | Yes | Yes | No | No | No | No |
| Qwen3.5-2B | 2026-03 | 262k | 2B | Yes | Yes | No | No | No | No |
| Qwen3.5-0.8B | 2026-03 | 262k | 800M | Yes | Yes | No | No | No | No |
| Qwen3.5-27B | 2026-02 | 262k | 27B | Yes | Yes | Yes | Yes | Yes | Yes |
| Qwen3.5-35B-A3B | 2026-02 | 262k | 35B | No | No | Yes | Yes | Yes | Yes |
| Qwen3.5-122B-A10B | 2026-02 | 262k | 122B | Yes | Yes | Yes | Yes | Yes | Yes |
| Qwen3.5-Flash | 2026-02 | 1m | — | Yes | Yes | No | No | No | No |
| Qwen3.5-397B-A17B | 2026-02 | 262k | 397B | Yes | Yes | Yes | Yes | Yes | Yes |
| Qwen3.5-Plus | 2026-02 | 1m | — | Yes | Yes | No | No | No | No |
| Qwen3.5-4B-Instruct | 2025-11 | 256k | 4B | Yes | Yes | No | No | No | No |
| Qwen3.5-32B | 2025-11 | 256k | 32B | Yes | Yes | No | No | No | No |
| Qwen3.5-32B-Instruct | 2025-11 | 256k | 32B | Yes | Yes | No | No | No | No |
Available From(6 providers)
Pricing
Popular comparisons in this family
- Qwen3.5-122B-A10B vs Qwen3.6-27B995
- Qwen3.5-35B-A3B vs Qwen3.6-27B899
- DeepSeek V4 Pro vs Qwen3.5-397B-A17B783
- Qwen3.5-122B-A10B vs Qwen3.6-35B-A3B755
- Qwen3.5-27B vs Qwen3.6-27B655
- DeepSeek V4 Flash vs Qwen3.5-27B571
- Qwen3.5-27B vs Qwen3.6-35B-A3B558
- DeepSeek V4 Flash vs Qwen3.5-397B-A17B552
- DeepSeek V4 Flash vs Qwen3.5-122B-A10B511
- Llama 3.1 Swallow 8B Instruct vs Qwen3.5-9B318
Frequently Asked Questions
- What is Qwen3.5 used for?
- Qwen3.5 is used for agents, text, and vision and multimodal work. The family description and listed model capabilities point to those workloads as the best fit.
- How does Qwen3.5 compare to Tongyi DeepResearch?
- Qwen3.5 by Alibaba is strongest where you need agents, while Tongyi DeepResearch by Alibaba is the closest related family to check for adjacent model selection. Qwen3.5 has 14 listed variants and reaches up to 1m context, while Tongyi DeepResearch reaches up to 131k context, so compare the specs and pricing tables before choosing a production model.
- Which Qwen3.5 model should I use?
- For the lowest listed input price, start with Qwen3.5-Flash through OpenRouter at $0.07/1M input tokens. For the most capable/latest local choice, evaluate Qwen3.5-27B with 262k context and reasoning, tool use, function calling, structured outputs, and multimodal inputs.
Related Model Families
Models(14)
Qwen-AgentWorld-35B-A3B
2026-06262k35B total, 3B active
ReasoningOpen Source
Qwen3.5-9B
2026-03262k9B3 providers
MultimodalOpen Source
Qwen3.5-4B
2026-03262k4B
MultimodalOpen Source
Qwen3.5-2B
2026-03262k2B
MultimodalOpen Source
Qwen3.5-0.8B
2026-03262k800M
MultimodalOpen Source
Qwen3.5-27B
2026-02262k27B4 providers
MultimodalReasoningOpen Source
Qwen3.5-35B-A3B
2026-02262k35B2 providers
ReasoningOpen Source
Qwen3.5-122B-A10B
2026-02262k122B3 providers
MultimodalReasoningOpen Source
Qwen3.5-Flash
2026-021m3 providers
MultimodalOpen Source
Qwen3.5-397B-A17B
2026-02262k397B4 providers
MultimodalReasoningOpen Source
Qwen3.5-Plus
2026-021m3 providers
MultimodalOpen Source
Qwen3.5-4B-Instruct
2025-11256k4B
MultimodalOpen Source
Qwen3.5-32B
2025-11256k32B
MultimodalOpen Source
Qwen3.5-32B-Instruct
2025-11256k32B
MultimodalOpen Source