Back to Blog

ModernBert: small new Retriever/Classifier workhorse, 8k context, 2T tokens,

ModernBert: small new Retriever/Classifier workhorse, 8k context, 2T tokens,

Answer.ai/LightOn released ModernBERT, an updated encoder-only model with 8k token context, trained on 2 trillion tokens including code, with 139M/395M parameters and state-of-the-art performance on retrieval, NLU, and code tasks. It features Alternating Attention layers mixing global and local attention. Gemini 2.0 Flash Thinking debuted as #1 in Chatbot Arena, and the O1 model scored top in reasoning benchmarks. Llama downloads surpassed 650 million, doubling in 3 months. OpenAI launched desktop app integrations with voice capabilities. Figure delivered its first humanoid robots commercially. Advances in robotics simulation and a new physics engine Genesis claiming 430,000x faster than real-time were highlighted.

Read original post

Turn insight into implementation

Want help turning this idea into a production system?

xAGI Labs helps teams scope, build, and deploy AI products, agent workflows, voice systems, and enterprise rollouts.

If this topic is relevant to your roadmap, we can translate "ModernBert: small new Retriever/Classifier workhorse, 8k context, 2T tokens, " into a concrete build plan and launch path.