Qwen-SEA-LION-v4.5-27B-IT
Qwen-SEA-LION-v4.5-27B-IT is a 27-billion-parameter multimodal model based on Qwen3.6-27B. It uses both linear and full attention. AI Singapore distills it from Qwen3.5-397B-A17B using SEA-Instruct-2602 for regional adaptation.
How this release works
The card describes configurable thinking, retained reasoning context and code work across a repository. Its example disables thinking with the chat-template setting enable_thinking=false. It explicitly says the older Qwen3 /no_think text switch does not provide the same control here.Qwen-SEA-LION-v4.5-27B-IT official model card
How to assess it for your work
Practical guidance. Choose this model for a comparison that combines a local-language instruction with an image or a code change. Record the thinking setting as part of each result. This helps distinguish a model-quality difference from a configuration difference. Inspect the proposed tool arguments before allowing an application to execute them.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | aisingapore/Qwen-SEA-LION-v4.5-27B-IT |
| Input and output | Text and image inputs; text outputQwen-SEA-LION-v4.5-27B-IT official model card |
| Context stated by the card | 262K tokensQwen-SEA-LION-v4.5-27B-IT official model card |
| License declared by the model card | MITQwen-SEA-LION-v4.5-27B-IT official model card |
| Post-training | Logit distillation and model mergingQwen-SEA-LION-v4.5-27B-IT official model card |
Related models
- Gemma-SEA-LION-v4.5-E2B-IT
AI Singapore · SEA-LION v4.5
A Gemma 4 E2B adaptation for regional languages with text, image, video and audio inputs.
Launch date not confirmedDownloadable - Qwen-SEA-LION-v4-32B-IT
AI Singapore · SEA-LION v4
A 32B Qwen3 text model with continued pretraining on Southeast Asian languages and regional instruction tuning.
Launch date not confirmedDownloadable - Qwen-SEA-LION-v4-8B-VL
AI Singapore · SEA-LION v4
An 8B Qwen3-VL model adapted with about nine million regional instruction-text pairs.
Launch date not confirmedDownloadable