Salesforce unveils Koa, its first reasoning model built on Nvidia’s Nemotron

Salesforce unveiled Koa, its first reasoning model, built on Nvidia's Nemotron-3-Super-120B foundation model and trained with Group Relative Policy Optimization (GRPO) for multi-turn reasoning and tool-use tasks inside Salesforce's Agentforce platform. Koa outperforms its Nemotron base model on multi-turn tool-use benchmarks and beats proprietary baselines on agentic reasoning and CRM-specific tasks, though Salesforce said it does not match frontier models from the largest labs. The model stems from a Salesforce-Nvidia collaboration announced in March 2026 and targets sales, marketing, and customer support workflows, with governance integration through Slack and Data 360 for regulated industries such as financial services and healthcare.

™/®Salesforce.com Salesforce unveils Koa, its first reasoning model built on Nvidia’s Nemotron The enterprise AI giant's custom model targets sales, marketing, and customer support workflows through a novel reinforcement learning approach Salesforce https://cryptobriefing.com/markets/salesforce/ just entered the custom reasoning model race with Koa, a purpose-built AI system derived from Nvidia https://cryptobriefing.com/markets/nvidia/ ’s Nemotron-3-Super-120B foundation model. The model is designed to handle multi-turn reasoning and tool-use tasks inside Salesforce’s Agentforce platform, with a laser focus on the kind of work that actually pays the bills: sales pipelines, marketing campaigns, and customer support tickets. What Koa actually does Koa is not a general-purpose chatbot. It’s a specialized reasoning model trained to navigate the complex, multi-step workflows that enterprise CRM systems demand. The model was trained exclusively on public and synthetically generated data. The training process uses a reinforcement learning technique called Group Relative Policy Optimization, or GRPO, which incorporates what Salesforce describes as a simulation-to-reward pipeline. In practical terms, the system learns by simulating real Agentforce workflows called Agent Scripts and optimizing for outcomes that mirror successful task completion. On benchmarks, Koa outperforms the Nemotron base model it was derived from, particularly in multi-turn tool-use scenarios. It also beats strong proprietary baselines across several metrics related to agentic reasoning and CRM-specific tasks. That said, it doesn’t match the performance of frontier models from the biggest labs. AI, tech, and the markets they move—in one daily briefing. Daily. Free. Join 34,000+ readers across crypto, finance, and policy. The Nvidia partnership powering the build Koa is the product of a Salesforce-Nvidia collaboration announced in March 2026, which integrated Nvidia’s Nemotron models into the Agentforce platform. The Nemotron-3-Super-120B model that serves as Koa’s foundation features a 1-million-token context window and uses a Mixture-of-Experts architecture, a design pattern where only a subset of the model’s parameters activate for any given task. The partnership also emphasizes governance capabilities for regulated industries, with integration through Slack and Data 360 context addressing data-residency requirements and compliance frameworks in sectors like financial services and healthcare. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .