David Vs. Goliath: Why Small Language Models Are Quietly Winning Where It Matters Most

In enterprise AI, the models that win won't be the largest. They'll be the ones that know exactly what they're built for.​​ ...

Google updates Android Bench with new LLMs, but Gemini still lags behind

Google created a benchmark earlier this year to evaluate how LLMs perform in Android app development, and Android Bench is ...
Tech Times

Tencent’s Hy3 Tests Whether Smaller AI Models Can Rival Larger Flagships

Tencent has released Hy3, the official version of its latest Hunyuan large language model, with a launch pitch built around a ...

Meta's next model 'Watermelon' matches GPT-5.5 performance: Report

According to the report, Watermelon is still in training and is expected to succeed Meta's previous frontier AI model, ...

Why LLM Leaderboards No Longer Decide Enterprise AI Buying

Databricks' benchmark suggests enterprise AI buyers are moving beyond public leaderboards, prioritizing real-world ...

Meta enters AI image model race in bid to court advertisers and subscribers

Meta has announced Muse Image, its first AI model for image creation, as it seeks to attract creators and advertisers to its ...

How DeepSeek’s DSpark Makes LLMs 85% Faster Without Retraining

DeepSeek has introduced DSpark, an open-source speculative decoding method that accelerates LLM inference by up to 85% ...
TMCnet

Workato Contributes Real-World Agentic Workflows to the First MLPerf® Agentic Inference Benchmark

Workato, the leading Control and Execution Platform for Enterprise AI, today announced its contribution to the MLPerf Agentic Inference Benchmark, the industry-standard AI benchmark suite from ...

Medical AI may look less biased on paper but not in practice, new study finds

Large language models (LLMs) are only as good as the data they learn from. If their training data contains social biases, the ...

SpaceXAI launches Grok 4.5, claims faster performance and lower costs than rival AI models

SpaceXAI has introduced Grok 4.5, its first major AI model since becoming a public company, claiming improved speed, lower pricing and higher token efficiency. The launch comes just before OpenAI is ...
The Next Web

The Eval Stack: Proving the agents are right instead of claiming It

Sixtyfour CEO Saarth Shah explains how his company built AI research agents around rigorous evaluation systems rather than language model fluency, creating verification infrastructure that proves ...