Hugging Face releases Olmo-core 3 open MoE training framework

Hugging Face released Olmo-core 3, an open training framework that scales mixture-of-experts models to the trillion-parameter range while preserving computational efficiency. In a benchmark, the expert pool grew from 8 to 128 and total parameters from 4.6B to 47B, with training throughput falling by less than 5%.

• Cloudflare open-sources Clef decision models, launches RL fine-tuning2026-10-01
• llama.cpp merges MTP support for Qwen Flash Next2026-10-01
• Is sandboxing sufficient to contain rogue agents?2026-10-01
• Hugging Face open-sources fastest WebGPU kernels for local AI2026-09-30
• Hugging Face launches open TTS leaderboard2026-09-30