2025-03-17
Despite the name implying it—this is not a GPT 4.5 level model. The second tweet in the thread compares ERNIE-4.5 to GPT-4.5 across 15 benchmarks and shows that ERNIE-4.5 scores an average of 79.6 vs 79.14 for GPT-4.5. Problem is the benchmarks in the average are cherry-picked.
Bloomberg
Baidu releases Ernie X1, an AI model that articulates its reasoning similarly to DeepSeek R1, and upgrades its flagship foundation model to Ernie 4.5, both free
And Costs Half As Much: ‘Excellent Multimodal Understanding Ability’ Nisha Gopalan / Yahoo Finance : China's Baidu Takes on DeepSeek With New AI Model Mike Wheatley / SiliconANGLE ...
2025-03-16
Despite the name implying it—this is not a GPT 4.5 level model. The second tweet in the thread compares ERNIE-4.5 to GPT-4.5 across 15 benchmarks and shows that ERNIE-4.5 scores an average of 79.6 vs 79.14 for GPT-4.5. Problem is the benchmarks in the average are cherry-picked.
Bloomberg
Baidu releases Ernie X1, an AI model that articulates its reasoning similarly to DeepSeek R1, and upgrades its flagship foundation model to Ernie 4.5, both free
The Ernie X1 model by China's internet search leader works similarly to DeepSeek R1 — which shocked Silicon Valley …