Faster AI Performance
Anthropic Launches Claude Sonnet 5.5 Featuring Faster Performance and Reduced Task Costs
Anonymous
Anthropic has launched Claude Sonnet 5.5, an update to its mid-tier model for everyday work including coding, debugging, document and spreadsheet creation, presentations, and interface design. The company says it generates output more than 30% faster than Sonnet 5 and can reduce task-completion costs by up to 30%, largely through fewer tokens and tool calls, while per-token API prices remain unchanged. Anthropic says Sonnet 5.5 performs particularly well on agentic coding and approaches its flagship Opus 5.5 on several evaluations, outperforming it on one reported command-line coding benchmark; results vary by workload. The release also brings clearer writing and cyber safeguards similar to those for higher-tier models, while Anthropic positions Opus for more open-ended work. Anthropic says a new Haiku model is planned for the coming weeks.
Key Points
1
Sonnet 5.5 includes invisible text watermarking intended to make AI-generated writing easier to detect without affecting readability; Anthropic said the feature helps meet regulations including the EU AI Act.
2
Zendesk’s director of AI said that, in tests involving hundreds of support replies and escalation requests, Sonnet 5.5 made fewer incorrect decisions and resolved tickets faster than the Claude models Zendesk currently uses in production.
3
On Terminal-Bench 4.0, Sonnet 5.5 scored 70.6%, compared with 66.4% for Opus 5.5 and 10.3% for Sonnet 5. The benchmark’s maintainers cautioned that timeouts and token limits contributed to Sonnet 5’s low result.
4
VentureBeat reported that Sonnet 5.5’s API pricing also includes cache writes at $2.50 per million tokens; Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, with cache writes priced at $5 per million tokens.
Sentiment
Mostly PositiveNEGATIVENEUTRALPOSITIVE
Publishers
36
Articles
24
Reach
60
A
AnonymousAuthor
