Terminal-Bench 3.0 has a new leader: Claude Opus 5 Max + mini-SWE-agent tops the board at 43.5%, beating GPT-5.6 Sol Max with Codex. But the benchmark measures the model and agent together, not the model alone.
Terminal-Bench 3.0 has a new leader: Claude Opus 5 Max + mini-SWE-agent tops the board at 43.5%, beating GPT-5.6 Sol Max with Codex. But the benchmark measures the model and agent together, not the model alone.
An OpenAI product lead showed developers how to swap GPT into Claude Code's harness. One user got banned for it, and the aftermath turned into a very public — and very petty — cross-company spat.
Herdr, a tmux-like tool built for managing coding agents, has joined Y Combinator's F26 batch. The solo-developed open-source project racked up 25K GitHub stars and 340K downloads in four months.
Hold the Fn key and speak: Gemini for macOS now transcribes and edits text in any app, cleaning up filler words and even rewriting content on command.
A Claude Max 20x pricing bug allowed free access to a $200/month plan, leading to mass account bans that also swept up paying customers and long-time users.
OpenAI reveals two cases where its internal models broke rules after hours of continuous work, warning developers about new risks from long-horizon tasks.
OpenAI's strategy lead Dean W. Ball says China's Kimi K3 model is so strong that it rivals top 2026 models—and warns the US may need to block it from American businesses.
Ryan Beiermeister, fired from OpenAI after opposing ChatGPT's adult mode, joins Peter Thiel's Founders Fund as an investment partner.
Mira Murati's Thinking Machines Lab releases Inkling, a 975B-param MoE model that uses a fraction of tokens compared to Nvidia's Nemotron.
