Claude Sonnet 5.5

Anthropic released Claude Sonnet 5.5 on September 28, 2026, as the second model in its Claude 5.5 family, positioned as a faster, lower-cost complement to Opus 5.5 for well-scoped everyday tasks like bug fixing and document creation. On Terminal-Bench 4.0, an agentic coding evaluation, Sonnet 5.5 scores 70.6% compared to Sonnet 5's 10.3%, and it is the first Sonnet to beat Pokémon Red working only from screenshots. Pricing matches its predecessor at $2 per million input tokens, $10 per million output tokens, and $0.20 per million tokens for cache reads, yet Sonnet 5.5 generates outputs 30%+ faster than Sonnet 5 and typically costs up to 30% less per task. It also launches with cyber safeguards modeled on the most capable Claude variants, with Claude Haiku 5.5 slated to join the family in the coming weeks.

The Decoder reports Anthropic released Claude Sonnet 5.5 on September 28, 2026, the second model in the Claude 5.5 family, generating output more than 30% faster and reducing per-task costs by up to 30% via more efficient token usage. The model nearly matches Opus 5.5 on several benchmarks at a fraction of the cost, targeting well-defined tasks like debugging and building presentations. Sonnet 5.5 is available on AWS, Google Cloud, and Azure at launch, and Anthropic added new cybersecurity and distillation safeguards for this release. On Terminal-Bench 4.0, it reaches 70.6% versus Sonnet 5's 10.3%, and on CursorBench 4.0 it scores 55.5%. Claude Haiku 5.5 is announced for the coming weeks to address high-throughput, low-cost use cases.

Vellum explains that Anthropic released Claude Sonnet 5.5 on September 28, 2026, six days after Claude Opus 5.5, positioning the mid-tier as a high-speed workhorse that runs more than 30% faster than Sonnet 5 and cuts effective per-task costs by up to 30%. On Terminal-Bench 4.0, Sonnet 5.5 jumps from Sonnet 5's 10.3% to 70.6%, edging past Opus 5.5's 66.4%. On GDPval-AA v2.1, Sonnet 5.5 scores 1844 Elo, two points shy of Opus 5.5 and roughly 360 points ahead of OpenAI's GPT-6 Sol. List pricing stays at $2 per million input tokens, $10 per million output tokens, and $0.20 per million prompt cache reads, with cost-per-task reductions driven by fewer steps and more aggressive tool batching.

VentureBeat reports Anthropic announced Claude Sonnet 5.5 on September 28, 2026, a mid-tier update that generates output more than 30% faster than Sonnet 5 and can cut total task cost by up to 30% through fewer tokens and fewer tool calls, not lower API list prices. Pricing holds at $2 per million input tokens and $10 per million output tokens, with $0.20 per million cache-read tokens. The model nearly matches Opus 5.5 on several evaluations, scoring 1844 on GDPval-AA versus Opus 5.5's 1846 and well ahead of Sonnet 5's 1449, while reaching 80.1% on OSWorld 2.1. Anthropic positions Sonnet 5.5 for well-defined everyday work while reserving Opus 5.5 for ambiguous, judgment-heavy tasks.

CodeRabbit evaluated Claude Sonnet 5.5 on real code-review workloads, finding it catches more bugs than Sonnet 5 at roughly half the wall-clock time and about 60% lower model-call cost per review. On 13 known-bug cases, Sonnet 5.5 caught 6 issues versus 4 for Sonnet 5, with 41.2% actionable precision versus 40.0% for Sonnet 5. The team rates Sonnet 5.5 as the first Sonnet release suitable for the main review pass rather than only for comment-quality refinement, noting it produces cleaner comments and resolves a broader share of latent bugs. A larger run across 44 open-source pull requests confirmed the speed and comment-volume gains.

Vals AI lists Claude Sonnet 5.5 as released September 28, 2026 by Anthropic, with a 1M context window, 128k max output tokens, and pricing of $2.00 input and $10.00 output per million tokens. The model supports text, image, and file input modalities, with video not supported and weights remaining private. Across Vals AI's independent benchmarks, Sonnet 5.5 achieves a Vals Index of 69.22% (±0.96), ranking second among 66 models, and leads on Code Migration (69.83%), Vibe Code Bench v1.1 (92.39%), and BioMysteryBench (81.11%). Its cost per test on the Vals Index is $20.80 with 2.27% fallback rate and 0.14% refusal rate.

DataCamp analyzes Claude Sonnet 5.5 as Anthropic's new mid-tier model, launched six days after Opus 5.5, replacing Sonnet 5 at the same price. The mid-tier closes most of the flagship's performance gap at roughly half the per-token price, reflecting logarithmic scaling where most capability gains remain relatively inexpensive to obtain. Key headline gains include agentic terminal coding where Sonnet 5.5 edges past Opus 5.5, and knowledge work plus computer use benchmarks where it sits just behind Opus 5.5. The article recommends making Sonnet 5.5 the default Claude model while reserving Opus 5.5 for open-ended, judgment-heavy work.