AI Briefing#openbmb
The latest AI and developer news about #openbmb, with the original source and a short summary.
The latest AI and developer news about #openbmb, with the original source and a short summary.
The agent optimized kernels against a black-box baseline without source code access, releasing the harness and 27 operators.
The 1B OCR model runs locally on mobile devices while the 2B verifier model checks output accuracy against original images.
The system uses Lookahead Expert Scheduling to preload weights, achieving 6.07x speedup over the baseline with only a 0.16-point benchmark drop.
The new 35B-A3B MoE model utilizes Lookahead Expert Scheduling to enable local vision-language processing on smartphones.
The new model achieves 13+ tokens/s decoding speed on iPhone 17 Pro Max using Lookahead Expert Scheduling.
OpenBMB has released MiniCPM5-2B, which achieved the highest intelligence index score among open-weight models with 4B parameters or fewer.
OpenBMB has released MiniCPM5-2B, a model that recorded top-tier performance among 4B-class models.
OpenBMB has released an open-source Embodied AI model series for robot manipulation and tracking.
Huawei unveiled the BitCPM-CANN model series, which supports 1.58-bit quantization-aware training on Ascend NPUs.
OpenBMB released BitCPM4-CANN 1B, 3B, and 8B on Hugging Face.