Home
Baidu Launches Ernie 5.1, Slashing Pre-training Costs by 94% and Ranking Among Top Four Globally
Baidu officially launched its next-generation language model, Ernie 5.1, on May 11, 2026. Building upon the Ernie 5.0 pre-training foundation released earlier in January—a model with 2.4 trillion parameters—this new version was developed using an innovative "one-shot elastic training framework." This approach enabled Baidu to optimize multiple model sizes in a single training cycle, slashing Ernie 5.1's pre-training costs to just 6% of comparable models.

As of May 9, Ernie 5.1 held the fourth position globally and ranked first among Chinese models on the Arena Search leaderboard with a score of 1223 points, showcasing its exceptional efficiency and well-balanced performance.
Architecturally, Ernie 5.1 features a sub-model design with tunable depth, width, and number of active experts. Its total parameter count is only one-third of its predecessor, and the number of effective parameters activated per query has been roughly halved. To address the "see-saw effect" common in multi-skill training, Baidu implemented a four-phase post-training process. This method utilizes parallel specialized training code, inference, and proxy expert models, combined with strategy distillation and reinforcement learning, effectively resolving the industry-wide challenge of balancing programming capability with creative reasoning. Furthermore, a revamped reinforcement learning infrastructure decouples model updates, response generation, and evaluation. When paired with a standardized low-precision computation library, this significantly enhances the stability of large-scale training.

Related article
Anthropic Revenue Surges, Market Cap Nears Trillion-Dollar Mark
Reports indicate that Anthropic has achieved substantial revenue growth in just a few months. The company's annualized revenue now exceeds $30 billion—more than triple what it was at the end of last year. This growth could give Anthropic a slight edg
Apple removes Cal AI app for unauthorized in-app purchases and manipulative billing
Apple’s recent removal of Cal AI, a popular AI-powered food tracking app within MyFitnessPal, underscores its strict enforcement of App Store policies on external payments and subscriptions. The app, which generates $50 million in annual recurring re
Github Copilot's token-based billing sparks developer outrage
The golden era of Microsoft's GitHub Copilot may be ending, especially for individual users. The company is shifting from a flat subscription fee to a token-based billing model, which could significantly increase costs. While larger enterprises might
Related Special Topic Recommendations
Comments (0)
0/500
Baidu officially launched its next-generation language model, Ernie 5.1, on May 11, 2026. Building upon the Ernie 5.0 pre-training foundation released earlier in January—a model with 2.4 trillion parameters—this new version was developed using an innovative "one-shot elastic training framework." This approach enabled Baidu to optimize multiple model sizes in a single training cycle, slashing Ernie 5.1's pre-training costs to just 6% of comparable models.

As of May 9, Ernie 5.1 held the fourth position globally and ranked first among Chinese models on the Arena Search leaderboard with a score of 1223 points, showcasing its exceptional efficiency and well-balanced performance.
Architecturally, Ernie 5.1 features a sub-model design with tunable depth, width, and number of active experts. Its total parameter count is only one-third of its predecessor, and the number of effective parameters activated per query has been roughly halved. To address the "see-saw effect" common in multi-skill training, Baidu implemented a four-phase post-training process. This method utilizes parallel specialized training code, inference, and proxy expert models, combined with strategy distillation and reinforcement learning, effectively resolving the industry-wide challenge of balancing programming capability with creative reasoning. Furthermore, a revamped reinforcement learning infrastructure decouples model updates, response generation, and evaluation. When paired with a standardized low-precision computation library, this significantly enhances the stability of large-scale training.

Anthropic Revenue Surges, Market Cap Nears Trillion-Dollar Mark
Reports indicate that Anthropic has achieved substantial revenue growth in just a few months. The company's annualized revenue now exceeds $30 billion—more than triple what it was at the end of last year. This growth could give Anthropic a slight edg
Apple removes Cal AI app for unauthorized in-app purchases and manipulative billing
Apple’s recent removal of Cal AI, a popular AI-powered food tracking app within MyFitnessPal, underscores its strict enforcement of App Store policies on external payments and subscriptions. The app, which generates $50 million in annual recurring re
Github Copilot's token-based billing sparks developer outrage
The golden era of Microsoft's GitHub Copilot may be ending, especially for individual users. The company is shifting from a flat subscription fee to a token-based billing model, which could significantly increase costs. While larger enterprises might











