Post Thumbnail

Salesforce test: Gemini 2.5 Pro solves only 58% of business tasks

The Salesforce CRMArena-Pro test shows that even leading artificial intelligence models face serious limitations when solving everyday business tasks.

Imagine: the flagship model Gemini 2.5 Pro successfully handles only 58% of requests with a single query. And what happens with multi-step dialogue? Efficiency rapidly drops to 35%!

CRMArena-Pro tests large language models under real conditions of sales, customer service and pricing. Researchers created 4280 unique tasks across 19 types of business operations using synthetic Salesforce data.

Particularly revealing are the results in multi-step dialogues — a key element of any business interaction. Almost half of Gemini 2.5 Pro’s failed attempts are related to inability to request critically important information. Models that ask more clarifying questions demonstrate significantly better results.

The highest performance was achieved in automating simple workflows — 83% success in routing support service requests. However, tasks requiring deep text understanding or following complex rules remain a serious challenge for modern artificial intelligence technologies.

Autor: AIvengo
For 5 years I have been working with machine learning and artificial intelligence. And this field never ceases to amaze, inspire and interest me.

Latest News

How Robomart reduces delivery costs by 70% through robotics

$3 for any delivery. Robomart challenges giants DoorDash and Uber Eats with a business model new to the industry. Their new robot RM5 completely changes delivery economics.

Unusual collaboration between competitors in AI safety testing

Two main competitors in the world of artificial intelligence united for the first time for joint safety testing. OpenAI and Anthropic opened access to each other's secret models. In an industry where companies pay researchers up to $100 million and fight for every user, such collaboration seems incredible.

Why Gemini reached 50% of ChatGPT's mobile audience

Google Gemini already has half of ChatGPT's audience on mobile devices. This is data from a new report by venture fund Andreessen Horowitz on the consumer AI market. 2.5 years of research shows an interesting picture.

How Claude became a hacking tool for 17 organizations

Anthropic company released an analytical security report. From it becomes clear that Claude and other AI agents are becoming tools of cybercriminals. At Anthropic, they called this new direction vibe-hacking. It turns out that artificial intelligence has radically lowered barriers to entry into criminal activity.

How xAI competes with OpenAI in developer tools

xAI is launching Grok Code Fast 1. This is a compact agentic model for coding. $0.20 for 1 million input tokens, $1.50 for output — and just $0.02 when using cache!