AI research firm DeepGrove has announced Maple-Preview, a new AI model designed for efficient operation on mobile devices like the iPhone. This model boasts 13 times the processing speed of Bonsai 27B, another iPhone-compatible AI, while maintaining comparable performance. Unlike traditional quantization methods that reduce AI model accuracy, Maple-Preview was engineered from the ground up using ternary processing to achieve both high performance and efficiency. The model is open-source, released under the MIT License, and available on Hugging Face, with a demo site also provided. AI
IMPACT Accelerates the trend of powerful AI models running directly on consumer devices, potentially increasing privacy and reducing reliance on cloud infrastructure.
RANK_REASON New model release from a research firm with performance claims and open-source availability. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Read on Mastodon — mastodon.social →
- Bonsai 27B
- DeepGrove
- Gemma 4 E4B
- GLM-4.7-Flash
- gpt-oss-20b
- iPhone
- Maple-Preview
- MIT License
- Qwen3.5 35B-A3B
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
