Corrections
When we get something wrong we fix it and note the change here and on the item.
Update · · OpenAI Alleges Ex-Employee Took Unreleased Device Files to Apple
We added a later development: at an Oct. 14 hearing, after the judge said he was highly likely to grant Apple a preliminary injunction, OpenAI voluntarily agreed not to use the information Apple says was taken while the case proceeds and to collect forensic images of devices from 14 employees, as Bloomberg reported. This is not a court finding against OpenAI.
Correction · · Google Releases Gemini 4 Argon, Its First Major Frontier Model Since Gemini 3
We labeled Gemini 4 Argon’s #1 Vals Index ranking as an unverified Google claim. Vals AI, which runs the index, lists Argon first on its own leaderboard (68.90%), with Claude Sonnet 5.5 second (67.04%), so we relabeled the ranking as independently checked and added Vals’ figures.
Update · · Meta Launches Muse Wonder: Claims Frontier Models for Personal and Pro Agents
We added independent benchmark results published after launch: Artificial Analysis Intelligence Index scores of 56 (Muse Wonder Full) and 52 (Muse Wonder), and Vals Index ranks of #5 (64.57%) and #10 (61.06%). We also noted that Meta has published the Muse Wonder model card.
Correction · · Z.ai Launches GLM-Legal, an AI Agent Model Built for Real Law-Firm Work
We said Z.ai’s LegalBench claim hadn’t been independently checked. In fact, Vals AI independently evaluated GLM-Legal on LegalBench (86.2%, #14 of 150 models). We also updated its Harvey LAB rank from #6 of 77 to #7 of 79 after Vals added newer models, and noted that Vals lists the model as open-weight.
Update · · Grok, Cursor and X Premium Merge Into One Plan: “The X Subscription”
We had labeled the $8–$200 plan prices as reported, not official. SpaceXAI’s announcement page and pricing docs list official prices (Free $0, Plus $8, Premium $30, Super $100, Ultra $200 a month), so we replaced the reported figures with the official ones.
Spotted an error? Tell us.