Skip to main content

News Roundup 26

Created on August 21|Last edited on August 23
Microsoft unveils Phi-3.5: advanced LLM models for diverse AI tasks
Microsoft releases three new Phi-3.5 models, excelling in reasoning, multimodal tasks, and efficiency

Fine-tuning for GPT-4o now available with free training tokens
OpenAI launches fine-tuning for GPT-4o, offering customized AI models with free tokens until September 23

Neural Magic releases LLM Compressor library for efficient model deployment
Neural Magic's LLM Compressor enhances large model efficiency with unified compression techniques and improved inference speeds

Uber and Cruise to launch autonomous vehicle partnership
Uber and Cruise will integrate self-driving cars into Uber's platform next year

LongWriter pushes LLM output beyond 2,000 words
New methods allow long context LLMs to generate coherent outputs over 20,000 words