Anthropic has introduced Claude Sonnet 5.5, the second model in its Claude 5.5 family, positioning it as a faster and more affordable alternative to Claude Opus 5.5 for everyday tasks.
According to Anthropic, Sonnet 5.5 represents a clear upgrade over Claude Sonnet 5. The model runs 30% or faster and costs up to 30% less per task despite maintaining the same token pricing as its predecessor at $2 per million input tokens and $10 per million output tokens. The efficiency gains come from the model requiring fewer tokens to complete the same work.
Performance improvements are particularly notable in coding tasks. On the Terminal-Bench 4.0 agentic coding evaluation, Sonnet 5.5 scores 70.6% compared to Sonnet 5’s 10.3%. On the FrontierCode benchmark, it achieves a 52.1% score at maximum effort, 10 points higher than Sonnet 5 at the same setting and at roughly one-fifteenth of the cost per task. The model is also the first Sonnet version to beat Pokémon Red using only screenshots.
Beyond coding, Sonnet 5.5 shows gains in knowledge work. On GDPval-AA, which tests models across real-world tasks spanning 44 occupations and nine industries, Sonnet 5.5 scores nearly level with Opus 5.5 and approximately 400 points above Sonnet 5. Early testers noted improvements in collaboration, describing the model as a better partner for iteration work and highlighting its design capabilities, including the ability to create polished presentation decks that require minimal editing.
Anthropic positions Sonnet 5.5 as strongest at well-scoped everyday tasks, bug fixing, and document creation, with capabilities that complement rather than replace Opus 5.5, which remains focused on complex work requiring sustained judgment. The company notes that while Sonnet 5.5 matches or improves over Sonnet 5 on most alignment measures, Opus 5.5 still performs slightly better overall on alignment testing.
Notably, Sonnet 5.5 is the first Sonnet model to launch with cybersecurity safeguards similar to those deployed on Opus 5.5, though routine software development work remains unaffected. Claude Haiku 5.5, designed for high-volume and cost-sensitive applications, will join the Claude 5.5 family in the coming weeks.
Key facts
- Sonnet 5.5 runs 30% faster than Sonnet 5 and costs up to 30% less per task
- On Terminal-Bench 4.0 agentic coding evaluation, Sonnet 5.5 scores 70.6% versus Sonnet 5’s 10.3%
- The model scores nearly level with Opus 5.5 on GDPval-AA knowledge work benchmark
- Sonnet 5.5 is the first Sonnet model to include cybersecurity safeguards like Opus 5.5
- Early testers noted improvements in collaboration and design capabilities for creating presentations