With its first LLM benchmark, Salesforce wants to support companies in their search for the right LLM for their CRM systems.

Salesforce launches the world’s first LLM benchmark for CRM. This new benchmark gives companies the guidance they need to make smart decisions when evaluating generative AI models for their CRM systems. Based on key metrics such as accuracy, cost, speed, trust and security, companies can make an informed decision. Salesforce continues to keep pace with technological developments by adding to and continuously improving the benchmark.

LLM Performance

Salesforce’s new LLM Benchmark provides a comprehensive view of the performance of different LLMs based on four metrics: accuracy, cost, speed and trust, and security.

It is designed to evaluate common sales and service use cases such as prospecting, lead nurturing, and opportunity summaries and service cases. In addition, the benchmark includes a public leaderboard to help professionals decide which LLM best fits their CRM needs.

“As AI continues to evolve, business leaders say it’s important to find the right mix of performance, accuracy, accountability and cost to realize the full potential of generative AI and drive business growth,” said Silvio Savarese, EVP and Chief Scientist at Salesforce AI Research.

developments

Salesforce will continue to add new use cases to the benchmark and improve the assessment of LLMs. “Sophisticated LLM” will also be added to the benchmark soon.

“Salesforce’s new LLM Benchmark for CRM is a significant step forward for companies in the industry to evaluate their AI strategy. Not only does it provide clarity in implementing next-generation AI, but it can also accelerate time to value for CRM-specific use cases,” said Savarese.