Researchers introduced a paired fixed-content probe over 500 MMLU-Pro items to test how discourse-role labels such as Instruction:, Reference:, and Example: affect language model adoption of misleadin…
At MLSys 2026, Modular identified three key trends in AI inference, highlighted by keynotes on agentic kernel development and the need for "zero trust" verification to prevent benchmark cheating. Lido…
Alibaba’s Qwen3.7-Max autonomously optimized a kernel on unfamiliar hardware for 35 hours, making 1,158 tool calls and achieving a 10x speedup over the reference implementation. The model outperformed…
According to the article, AI API pricing in 2026 is highly fragmented, with a 300x difference between the cheapest and most expensive models; for example, a prompt costing $30 on GPT-5.5 costs only $0…
The article explains that OpenCode Go's usage limits are based on dollar amounts ($12 per 5-hour window) rather than request counts, making model routing critical because different models offer vastly…
Modular released AI agent skills for its Mojo language that enable coding assistants to translate existing GPU kernels from CUDA and Triton into Mojo code, addressing the challenge that large language…
New research shows frontier AI models have improved nearly tenfold in their ability to follow instructions over the past year, according to data from the IFScale benchmark. A year ago, models began lo…
Seven AI models were each given the same prompt in blind OpenCode sessions and asked to rank all seven, including themselves, on fixed factors. The mistakes — silent model substitutions, hallucinated …