Analyst recommends Opus 5.5 for most use cases despite Sonnet 5.5's price advantage
Claude Sonnet 5.5 release benchmarks
Latest: Claude Sonnet 5.5 Ranks #2 on AI Index, Near Opus 5.5 at Lower Cost
Sonnet 5.5 brings near-flagship benchmark performance at a mid-tier price point, which affects how teams choose between Anthropic models for agentic and production workloads. Independent benchmark scores diverge somewhat from Anthropic's own figures, and at least one real-world cost test shows savings exceeding Anthropic's stated range, so actual efficiency depends heavily on workload type.
When stories like this move, I write about them in an email each morning.
One free email each morning, around 8 AM US Eastern.
Check your email.
I have sent you a confirmation link. One click and you are in.
Not in your inbox? Look in spam or promotions. If it's there, mark it "Not spam" so future editions land in your inbox.
The full picture
Anthropic's Claude Sonnet 5.5 holds second place on the Artificial Analysis Intelligence Index, behind only Opus 5.5, with Artificial Analysis independently scoring it at 64% on Terminal-Bench 4.0 (max setting), a 50-point gain over Sonnet 5 (max) and slightly above Opus 5.5 and GPT-6 Astra on that benchmark. Anthropic's own published score for Terminal-Bench 4.0 is 70.6% versus Sonnet 5's 10.3%, and on GDPval-AA the model scores 1,844 against Opus 5.5's 1,846 and Sonnet 5's 1,449. The list price is unchanged from Sonnet 5 at $2/1M input and $10/1M output tokens; Anthropic claims effective per-task costs fall by up to 30 percent because the model completes tasks with fewer tokens and runs more than 30 percent faster. Third-party pipeline tests show even larger savings in some workloads: CodeRabbit found Sonnet 5.5 cost roughly 40 percent of Sonnet 5 for equivalent code reviews, a saving of about 60 percent per review. One analyst recommended sticking with Opus 5.5 for most use cases despite the price difference, and Anthropic itself states Opus 5.5 remains clearly stronger on complex open-ended work.
How it developed
Agentic test shows Sonnet 5.5 completing a multi-bug Python fix in 3 tool calls versus Sonnet 5's 12
Anthropic released Claude Sonnet 5.5 on October 5 at Sonnet 5's price of $2/$10 per million tokens, reporting Terminal-Bench 4.0 at 70.6% versus Sonnet 5's 10.3% and GDPval-AA at 1,844, two points below Opus 5.5.
Artificial Analysis independently placed the model second on its Intelligence Index, measuring 64% on Terminal-Bench 4.0 (max setting), below Anthropic's stated figure. CodeRabbit found API costs at roughly 40% of Sonnet 5 for equivalent code reviews.
Sources
6 more sources
Related
- Related storyClaude 5.5 family and GPT-6 Sol
- Anthropic's Claude Sonnet 5.5 release
- Anthropic makes Claude Sonnet 5's $2/$10 per million token pricing permanent
- Anthropic Launches Claude Haiku 5.5 at 75% Lower Average Cost
- Gemini 4 Carbon release
BackgroundHow a language model works: Tokens
Get the daily email
When stories like this move, I write about them in an email each morning.
One free email each morning, around 8 AM US Eastern.
Check your email.
I have sent you a confirmation link. One click and you are in.
Not in your inbox? Look in spam or promotions. If it's there, mark it "Not spam" so future editions land in your inbox.