Anthropic's mid-tier Claude 5.5 model: faster and cheaper than Sonnet 5, aimed at everyday coding and knowledge work.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Claude Sonnet 5.5 is Anthropic's second model in the Claude 5.5 family, released on September 28, 2026 as a faster, lower-cost complement to Claude Opus 5.5. Anthropic positions it for well-scoped everyday tasks, bug fixing, and producing documents, slides and spreadsheets, while Opus 5.5 remains the choice for complex, open-ended work requiring sustained judgment. Claude Haiku 5.5 was announced as coming later.
Anthropic reports 70.6% on Terminal-Bench 4.0 (Sonnet 5: 10.3%), 46.2% on FrontierCode 1.1 at Max effort, 55.5% on CursorBench 4.0, 80.1% on OSWorld 2.1 (partial), and GDPval-AA scores close to Opus 5.5. It runs 30%+ faster than Sonnet 5 and is described as the first Sonnet model to beat Pokémon Red from screenshots alone. It is also the first Sonnet launched with cyber safeguards and fallbacks, which target a narrow set of high-risk requests.
Pricing is $2 per million input tokens and $10 per million output tokens, with cache reads at $0.10 per million; Anthropic says it typically costs up to 30% less per task than Sonnet 5 because it uses fewer tokens. Effort levels (Low to Max) trade cost for score; Medium is the default in the Claude apps and High on the Claude Platform. Reviewers note that at Max effort it can use many output tokens and cost more per task than Opus 5.5, and that it is weaker at sound/music tasks.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Anthropic's new smaller model, close to Opus 5.5 on agentic coding and knowledge work, and now the default on the free plan. It ranks second on the Artificial Analysis index but uses the most output tokens per task, making it the most expensive to run.
Anthropic's cheaper model at $2/$10 per million tokens, tops Terminal Bench but trails Opus 5.5 elsewhere. Creator criticizes benchmarking at max effort, where it costs more per task than Opus 5.5 ($7.62 vs $5.98) and uses the most tokens. In his Beauty Bench it ranked 13th, behind GPT-6.1 Sol.
Creator demos Sonnet 5.5 building a 3D ocean, a Fall Guys clone, a Lego builder, an Unreal Engine San Francisco and other games from single prompts. Benchmarks are near Opus 5.5 (beats it on Terminal Bench 4.0) at half the price ($2/$10 per million tokens). Caveats: weak at sound/music, British spellings in writing.
Closest neighbours by meaning, not just by category.
Model · Anthropic
Anthropic's Claude model for coding, agents and long-running tasks, available via Claude apps, API and major clouds.
Model · Anthropic
Anthropic's mid-tier Claude model, used as a general-purpose workhorse for coding, agents and knowledge work.
Model · Anthropic
Anthropic's Claude Sonnet 5 LLM, with a high-effort "Max" setting; since succeeded by Claude Sonnet 5.5.
Model · Anthropic
Anthropic's Claude Opus 5.5 model for agentic coding and knowledge work, priced lower per task than Opus 5.
Model · Anthropic
Anthropic's Claude Opus 4.5 model for agentic coding, tool use and knowledge work, available via app and API.
Model · Anthropic
Anthropic's Claude Haiku 5.5, a small, fast and low-cost Claude model.
