Live--:--:-- PT

New Mexico Supreme Court fines lawyer $5,000 for ChatGPT-fabricated witnessesMoonshot AI targets $2B annualized revenue on Kimi K3 tractionMeta reworks AI prompt suggestions after chatbot probes user's childrenAnthropic details four cases of its own AI models hacking outside companiesTimnit Gebru says AI doom talk is a distraction from real harmsAnthropic blocks scientists from using Claude for bioweapon researchAI researchers quit frontier labs over extinction fearsAltman tells OpenAI staff the company is open to slowing AI development

Category

Models

Foundation models, benchmarks, and the research race.

News Business Tools Careers Security Faith Analysis

Models — page 8

OpenAI ships GPT-5.4 Thinking to all Plus users

The o-series-style reasoning model lands in ChatGPT with adaptive budgets — and a new opt-in 'extended thinking' toggle.

Jaeden SchaferApr 184 min read

Gemini 3.1 Pro sets a new reasoning high-water mark

Google's flagship now leads GPT-5.2 and Claude Opus 4.6 on GPQA Diamond, with 94.3% on a benchmark that has broken most prior models.

Jaeden SchaferApr 144 min read

Microsoft debuts three new foundation models

Text, voice, and image — Microsoft AI ships its own stack, ending years of quiet dependence on OpenAI for frontier capabilities.

Jaeden SchaferApr 135 min read