Koa vs. GLM-5.3

Salesforce Koa marks the company's first CRM reasoning model specifically designed for Agentforce, leveraging the power of NVIDIA Nemotron and drawing from an extensive 27 years of Salesforce CRM knowledge to adeptly handle complex, multi-step tasks in enterprise settings. This model is built upon nearly three decades of CRM application expertise and has been further refined through training on a specialized synthetic dataset that mirrors genuine business workflows, processes, and compliance requirements. The training scenarios are meticulously designed to emulate the reasoning, tool usage, and decision-making strategies that Agentforce agents utilize throughout the customer journey, including critical activities such as lead generation, opportunity qualification, and service case resolution across more than 14 distinct industries. Koa is purpose-built for precise CRM functionalities and is evaluated using the Salesforce CRM Bench, which includes authentic workflows like opportunity updates, case routing, and follow-up scheduling. Salesforce reports that Koa delivers an 11% improvement in accuracy for choosing the right actions and recalls customer context with a reliability that is 2.1 times superior to earlier models. This groundbreaking methodology not only boosts operational efficiency but also greatly enhances the overall customer experience, making it a vital tool for businesses aiming to optimize their customer interactions. By integrating advanced reasoning capabilities, Koa positions itself as a transformative solution in the realm of customer relationship management.

GLM-5.3 is Z.ai’s frontier coding model built to improve complex software engineering, long-horizon agent work, and advanced technical reasoning through scaled post-training. The model uses the same base model as GLM-5.2, with performance gains coming from additional post-training environments, more diverse tasks, and expanded compute on the existing training stack. Z.ai’s stack includes IndexShare for efficient long-context processing, SAO for reinforcement learning on long-horizon tasks, and slime for large-scale asynchronous post-training. GLM-5.3 is designed to perform better on work that resembles real engineering tasks rather than short coding exercises. Its training environments include production-style workflows where the model must diagnose bottlenecks, inspect documentation, use codebases, run experiments, implement changes, and produce measurable improvements. The model improves coding performance across public and private benchmarks, including Terminal Bench 3.0, DeepSWE, Agents’ Last Exam, and Z.ai Code Bench. GLM-5.3 also improves token efficiency, producing stronger agentic coding results than GLM-5.2 while using fewer output tokens in Z.ai’s internal evaluations. The model supports three reasoning effort levels, low, high, and max, and no longer supports disabling thinking. Z.ai recommends max reasoning effort for coding tasks, while applications using disabled thinking must migrate to enabled thinking before switching to GLM-5.3. The release also reports emergent cyber capabilities, including stronger vulnerability discovery and exploitation-chain reasoning, with open-weight release planned after safety evaluation and hardening. By combining scaled post-training, long-context infrastructure, long-horizon reinforcement learning, coding-agent workflows, benchmark improvements, reasoning controls, and ZCode integration, GLM-5.3 helps developers and researchers work on demanding coding and agentic tasks.