Google’s ToolGrad Turns Tool-Use Training Data Into an Answer-First Bet
Google Research's ToolGrad generates the answer before the...
Browse Tag
LLM Benchmarks
6 Articles
Five Startups Turn Transformer Compute Costs Into an Architecture Race
Five AI startups are betting that sparse attention, recurrent...
Google’s Science One Framework Turns AI-Generated Research Into a Verifiability Test
Google Research's Chain-of-Evidence standard and Science One...
SentinelOne’s Fast16 Benchmark Turns AI Malware Analysis Into a Recovery Test
SentinelOne built an eight-stage benchmark from a real...
NVIDIA’s Nemotron 3 Embed Turns Retrieval Accuracy Into an Agent Efficiency Problem
NVIDIA's new Nemotron 3 Embed flagship model ranks first on the...
Ai2’s olmo-eval Targets the Messy Middle of LLM Testing
Ai2 released olmo-eval, an open workbench for repeated LLM...