Records found: 4
NVIDIA's TiDAR combines one-step diffusion drafting with autoregressive verification in a single forward pass to exploit free GPU token slots and multiply tokens-per-forward by up to about 6x while preserving benchmark quality.
CoAct-1 introduces a hybrid approach combining GUI manipulation and direct coding to improve efficiency and reliability in autonomous computer operation, achieving a record 60.76% success rate on the OSWorld benchmark.
Liquid AI announces LFM2, an advanced edge AI model series delivering faster inference and training, with a novel hybrid architecture optimized for deployment on resource-constrained devices.
RWKV-X introduces a hybrid model combining sparse attention and recurrent memory to enable efficient decoding of extremely long sequences with linear complexity, outperforming previous RWKV models on long-context tasks.