The Lion-Shan AI lab under China Merchants' advanced technology institute released an embodied reasoning vision-language model called Lion-ER-35B-A3B, according to QbitAI on November 4. The mixture-of-experts model has about 35 billion total parameters and activates roughly 3 billion per token, improving on all ten mainstream embodied benchmarks and beating larger models on several. Its training set holds 312,380 samples, with budget allocated by capability gaps rather than data volume.
Source: public reports and official announcements
广告

