IBM's CEO says open AI models work 90% as well as top models. Per token, they can cost a tenth.
IBM CEO Arvind Krishna made the case on Wednesday's earnings call, The Register reported.
In 2024, he said, a top model might score 70 on a broad benchmark and an open model 20 or 30. By late 2025 the gap was more like 70 to 50.
Open models now cost 10% to 25% of the token price of the leading models, he said. "Would you like something that's 10 percent of the cost if it's 90 percent as good as the best, or do you want to pay that 10 times higher cost?"
Most tasks, such as summarising an email or examining a contract, can be done with the cheaper option, Krishna said. He expects more companies to run AI on their own hardware, which IBM sells.
The same call showed where IBM's own AI services stand. Finance chief Jim Kavanaugh said AI is about 30% of consulting revenue. Yet consulting revenue was flat at $5.34 billion, CNBC reported, while software grew 8% to $7.80 billion.
Our reading: Krishna's 90% is his own estimate. IBM sells both sides of his argument: the on-premise route to cheaper AI, and the consulting work where AI is now about 30% of revenue.
Sources
Our file on IBM
- 9 Oct
IBM sold the contracts of hundreds of VMware customers. It did not answer questions.
- 20 Sept
IBM and Nvidia still plan quantum computers. NEC will not build one.
- 16 Sept
KPMG and MNP put Quebec's $96 million health ERP above $1 billion.
For your next steering committee
- 1Ask which AI tools already touch your ERP or finance data, including the ones nobody approved.
- 2Ask what each tool is allowed to change, and who reviews what it changed.
- 3Ask how the result is measured, and who decides to switch a tool off.
Issue 01 of the ERP LEADERS brief puts a timeline, the three-number review and six questions on one page. Read issue 01
Welcome back. · Issue 01
