AIThe Deep View1h ago
Inception launches Mercury 3, a diffusion-based LLM it claims is
Inception launches Mercury 3, a diffusion-based LLM it claims is "comparable to Luna and Haiku models" but 4x to 8x faster, generating ~1,000 tokens per second
TL;DRNew AI model Mercury 3 matches competitor performance at 4-8x faster speeds using diffusion tech.
Why it matters: Faster inference dramatically reduces latency and compute costs, making advanced AI more accessible and practical for real-time applications.
Inception launches Mercury 3, a diffusion-based LLM it claims is “comparable to Luna and Haiku models” but 4x to 8x faster, generating ~1,000 tokens per second — W — hile the AI industry has bet heavily on LLMs and the massive data centers needed to scale them …
Read full articleSource: The Deep View · Opens in new tab