Gemma-SEA-LION-v4.5-E2B-IT

[Last update: 2026-06-17]

Gemma-SEA-LION-v4.5-E2B-IT is built upon the gemma-4-E2B-it architecture with 2.3B effective (5.1B with embeddings). To ensure deep domain adaptation, the model underwent distillation from google/gemma-4-31B-it on an updated aisingapore/SEA-Instruct-2602, instilling multilingual and multicultural fluency across English and key SEA languages: Burmese, Indonesian, Filipino (Tagalog), Malay, Tamil, Thai, and Vietnamese.

Gemma-SEA-LION-v4.5-E2B-IT inherits the following features from Gemma 4:

• Native System Prompt Support – Gemma 4 introduces native support for the system role, enabling more structured and controllable conversations.
• Reasoning – Highly capable reasoning model, with configurable thinking modes.
• Extended Multimodalities – Processes Text, Image with variable aspect ratio and resolution support (all models), Video, and Audio (featured natively).
• Optimized for On-Device – designed for efficient local execution on laptops and mobile devices.
• Enhanced Coding & Agentic Capabilities – Achieves notable improvements in coding benchmarks alongside native function-calling support, powering highly capable autonomous agents.

SEA-LION stands for Southeast Asian Languages In One Network.

We performed post-training in English and SEA languages on gemma-4-E2B-it, a multimodal learning model using the Gemma 4 architecture, to create Gemma-SEA-LION-v4.5-E2B-IT.

For tokenization, the model employs the default tokenizer used in gemma-4-E2B-it.

• Developed by: AI Products Pillar, AI Singapore
• Funded by: Singapore NRF
• Shared by: AI Products Pillar, AI Singapore
• Model type: Transformer Decoder with Vision and Audio Encoder
• Training Stage: Post-Training (Logit Distillation & Model Merging)
• Context length: 128k
• Language(s): fine-tuned on Burmese, Indonesian, Filipino, Malay, Tamil, Thai, and Vietnamese
• License: MIT
• Finetuned from model: https://huggingface.co/google/gemma-4-E4B-it

For more details, please refer to AI Singapore’s HuggingFace page for this model. The original GGUF files can be obtained from this HuggingFace repository.

Additional note: To disable thinking in your API calls to the model, submit "reasoning_effort": "none" under the parameters.

curl http://localhost:11434/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "aisingapore/Gemma-SEA-LION-v4.5-E2B-IT",
    "messages": [
      { "role": "user", "content": "Translate the following to Indonesian: I enjoy playing racquet sports such as tennis." }
    ],
    "stream": false,
    "reasoning_effort": "none"
  }'