The distilled version
- Anthropic released Sonnet 5.5, its mid-tier Claude model, saying it is 30 percent faster than Sonnet 5, burns tokens more slowly and costs significantly less.
- Anthropic pitches it as an everyday assistant for coding and office documents, and its own benchmarks show it beating the bigger Opus 5.5 model on agentic coding tasks.
- If you pay for Claude by usage (API or per-token billing), check whether your tools have switched to Sonnet 5.5 and whether your bill has dropped.
What happened
Anthropic released Sonnet 5.5, an update to its mid-tier Claude model, about three months after Sonnet 5 was announced. Anthropic says the new version is 30 percent faster than its predecessor and uses tokens (the units AI billing is based on) at a significantly slower rate, and that it will do the work for significantly less.
Anthropic describes Sonnet 5.5 as suited to everyday tasks including coding and creating office documents. The company’s own benchmarks show it outperforming its more powerful Opus 5.5 model on agentic coding tasks; TechCrunch suggests this is likely because Sonnet can run multiple agents at once without going over cost limits.
Anthropic also says Sonnet 5.5 has cyber capabilities comparable to its Opus 5 model, and that as a result it will be the first Sonnet version subject to the same cyber safeguards used for Anthropic’s Opus and Fable models. Anthropic plans to release an updated version of Haiku, its smallest model, in the coming weeks, but has not given a release date.
Why it matters
Any small business using Claude, directly or through a tool built on it, could see faster responses and lower costs on the same everyday work, such as drafting documents and writing code.
What it means for your business
If you run the business: ask whoever manages your AI tools whether they use Claude, and if so, whether they’ve moved to Sonnet 5.5. If pricing and speed are genuinely better as Anthropic claims, this is a free improvement with nothing to learn.
What to do
Act now- Check which Claude model your AI tools or API integrations currently use.
- If you pay by usage, decide whether to move to Sonnet 5.5, then compare next month’s costs with this month’s.
- Run a handful of your usual tasks (document drafts, coding, automation steps) on Sonnet 5.5 to see if speed and quality hold up.
- Watch for Anthropic’s coming update to Haiku, its cheapest model, if your tasks are simple enough to not need Sonnet at all.
Act now: worth doing this month.
Hype check
The claim: Anthropic calls Sonnet 5.5 significantly cheaper and faster than its predecessor.
The reality: Anthropic’s speed and cost figures come from the company itself, with no independent numbers given. The benchmark data in the report is separate: it shows Sonnet 5.5 beating Anthropic’s own Opus 5.5 on agentic coding, which is not evidence for the speed or cost claims.
Jargon in this story
Sources
- 1Anthropic releases Sonnet 5.5, which it calls a significantly cheaper, faster work partner (opens in a new tab) · TechCrunch · News report · Sep 28, 2026