Topics
Artificial intelligence and the technology around it: models, agents, tools and the companies behind them, explained in depth.
SpaceXAI released Grok 4.7 on 21 September 2026 at the same USD 2 and USD 6 per million tokens as Grok 4.6, live from day one inside Cursor. Independent tests rank it fourth among coding agents and found it uses more than twice as many output tokens per task.
Suno released its v6 model family on 9 September 2026, trained from scratch on a new dataset that includes licensed music, and retired all earlier models. A day earlier it signed Believe and TuneCore, which had refused to distribute music from Suno's older models.
Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 on 1 September 2026: one model under two sets of safeguards. The public Fable 5.1 may now find software bugs but not build exploits, and a 75 percent cut in cache read prices makes long agent runs cheaper at the same list price.
On 31 July 2026 a Munich court found that Suno copied six songs from GEMA's repertoire, kept them in its models and played them back to users. Commentators called the decision Europe's first AI music copyright ruling, and it reaches training that Suno did in the United States.
Anthropic released Claude Opus 5 on 24 July 2026 at USD 5 per million input tokens, the same as Opus 4.8 and half the price of Claude Fable 5, which it nearly matches on coding and agent tests. Anthropic expects its cyber filters to step in far less often, and it carries no 30 day data retention rule.
Anthropic released Claude Sonnet 5 on 30 June 2026 as the default model on Claude's Free and Pro plans, saying it comes close to Opus 4.8 on agent work. API prices opened at USD 2 and USD 10 per million tokens, while Haiku 4.5 stayed Anthropic's small model.
On 9 June 2026 Anthropic released Claude Fable 5, its first Mythos class model for the general public, and Claude Mythos 5, the same model with some safeguards lifted for vetted cyber defenders. Both cost USD 10 per million input tokens; what sets them apart is a set of classifiers.
At WWDC on 8 June 2026 Apple unveiled Siri AI, an assistant that reads personal context across apps and runs on Apple Foundation Models built with Google's Gemini. The same day Apple said EU iPhone and iPad owners would not get it at launch, blaming the DMA; Brussels said the choice was Apple's.
DeepSeek released preview versions of V4-Pro and V4-Flash on 24 April 2026 as open weights under the MIT licence, with 1 million tokens of context by default. Huawei said its whole Ascend supernode line supports them, a sign of a Chinese frontier model built around Chinese chips.
On 7 April 2026 Anthropic unveiled Claude Mythos Preview, its most capable model, and said it would not release it generally because it finds and exploits software flaws better than almost any human. About 50 organisations get it through Project Glasswing, for defence only.
There is a moment in every emerging technology where the line between "this is coming soon" and "this is already here" becomes impossibly blurry. Autonomous AI agents are at that line right now.
Your phone buzzes at 2:14 in the morning. It is not a person. It is an AI agent that has just finished reading 47 unread messages across three Telegram groups, cross-referenced them against its knowledge base, and drafted a precise response for your review.