Holo3.1 has been released, featuring enhanced robustness for local and mobile environments, quantized checkpoints for local inference, and improved performance across various deployment frameworks. This release addresses the challenges of deployment flexibility and performance consistency in diverse operational settings.
The Holo3.1 family has been introduced to provide seamless integration across desktop and mobile environments. It addresses users' desires for deployment flexibility and performance consistency across agent frameworks.
Holo3.1 improves robustness in three primary areas: environments, agent frameworks, and deployment targets.
Performance on mobile devices has significantly increased, with the 35B-A3B model improving from 67% to 79.3% on AndroidWorld.
This version includes quantized checkpoints optimized for local execution, such as FP8, Q4 GGUF, and NVFP4.
It introduces small model sizes (0.8B, 4B, and 9B) for cost-effective and private deployment.
New native support for function-calling protocols has been added, complementing existing JSON output formats.
Evaluations in various benchmarks like OSWorld and internal tests show that function-calling performance now approaches that of dedicated execution frameworks.
Holo3.1 improves operational efficiency across e-commerce, business software, and collaboration workflows.
The new model's capabilities make it easier for teams to integrate Holo3.1 into third-party agent stacks and realize significant performance gains.
β¨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors β check the original sources. How BrevFeed works β
Holo3.1 has been released, featuring enhanced robustness for local and mobile environments, quantized checkpoints for local inference, and improved performance across various deployment frameworks. This release addresses the challenges of deployment flexibility and performance consistency in diverse operational settings.