OpenAI employees publicly criticize the misleading Benchmark test results of the latest AI model Grok3 by XAI

ChainCatcher news, according to Jinshi report, an employee of OpenAI publicly accused Musk's XAI company of releasing misleading Benchmark test results for its latest AI model Grok3. In response, Igor Babushkin, co-founder of XAI, insisted that the company did nothing wrong. The chart of XAI shows that both versions of Grok3 - Grok3 Reasoning Beta and Grok3 mini Reasoning - outperform OpenAI's current strongest available model o3-mini-high on AIME 2025. However, OpenAI employees quickly pointed out on the X platform that the chart of XAI did not include the AIME 2025 score of o3-mini-high under the condition of cons@64. Babushkin argued on the X platform that OpenAI had also released similar misleading Benchmark test charts in the past, although these charts were used to compare the performance of its own models.