Image Credits:Ruhani Kaur/Bloomberg / Getty Images11:00 AM PDT · September 28, 2026
As the AI exemplary wars continue, Anthropic has released the newest mentation of Sonnet, the company’s mid-tier model, which it says volition enactment overmuch faster (and for importantly less) than its predecessor.
The laboratory describes Sonnet 5.5 arsenic an perfect adjunct for mundane tasks — including coding and creating bureau documents.
5.5’s predecessor, Sonnet 5, was announced astir 3 months ago. At the time, the model’s selling constituent was businesslike agentic deployment — the quality to tally agents astatine a little outgo than competitors.
The large selling constituent with 5.5, meanwhile, is speed. Anthropic claims that Sonnet 5.5 is 30 percent faster than its predecessor, and that its complaint of token pain is importantly slower.
In the Anthropic hierarchy of models, Sonnet is little almighty than the Opus model, but tin beryllium much utile successful definite circumstances due to the fact that its agility. In particular, Anthropic’s benchmarks amusement Sonnet 5.5 performing amended than Opus 5.5 connected agentic coding, apt due to the fact that of its quality to spawn aggregate agents without exceeding outgo limits.
Sonnet 5.5 is besides said to person important cyber capabilities, with the institution claiming that it has “comparable” cyber capabilities to Opus 5. As a result, Anthropic says that 5.5 is the archetypal Sonnet exemplary that volition beryllium taxable the aforesaid cyber safeguards that use to Fable and Opus.
The institution besides plans to merchandise a caller mentation of Haiku — its smallest exemplary — successful the coming weeks, though it didn’t springiness a steadfast day arsenic to erstwhile that would happen.
The past twelvemonth has seen a flurry of caller exemplary releases from the large AI labs. Just past week, OpenAI released a fig of caller models — including enhanced versions of Sol and Luna, its mid-tier and budget-friendly models. Meta besides announced a caller model, which it said would powerfulness an upcoming diagnostic associated with its astute glasses.
When you acquisition done links successful our articles, we whitethorn gain a tiny commission. This doesn’t impact our editorial independence.
Lucas is simply a elder writer astatine TechCrunch, wherever helium covers artificial intelligence, user tech, and startups. He antecedently covered AI and cybersecurity astatine Gizmodo. You tin interaction Lucas by emailing lucas.ropek@techcrunch.com.















English (US) ·