The short version
Sonnet 5.5 is Anthropic's new mid-tier model, positioned as a "faster, lower-cost complement to Claude Opus 5.5." Where Opus 5.5 targets complex work that needs careful judgment, Sonnet 5.5 is strongest at well-scoped everyday tasks — fixing bugs, and producing polished documents, slides, and spreadsheets. Anthropic also notes it "has a sharp eye for design"[1].
The practical headline is efficiency. Sonnet 5.5 keeps the same per-token price as Sonnet 5 but typically needs far fewer tokens to finish the same job, so it costs up to 30% less per task while running noticeably faster[1].
Performance: a big jump in agentic coding
The most striking improvement is in agentic coding. On Terminal-Bench 4.0, which measures complex multi-step professional tasks in a command-line interface, Sonnet 5.5 scores 70.6% against Sonnet 5's 10.3%[1]. It also scores nearly level with Opus 5.5 on real-world knowledge work[1]:
| Benchmark | Sonnet 5.5 | Sonnet 5 | Opus 5.5 |
|---|---|---|---|
| Terminal-Bench 4.0 | 70.6% | 10.3% | 66.4% |
| CursorBench 4.0 | 55.5% | 34.1% | 57.8% |
| FrontierCode 1.1 (Main) | 52.1% | 42.4% | 54.4% |
| GDPval-AA v2.1 | 1844 | 1449 | 1846 |
| Humanity's Last Exam | 64.5% | 54.9% | 67.7% |
| OSWorld 2.1 | 80.1% | 57.0% | 81.8% |
Anthropic-reported benchmark scores, September 2026[1]. FrontierCode shown at high effort.
Anthropic adds a fun data point that underscores the jump in long-horizon and image-understanding ability: Sonnet 5.5 is the first Sonnet model to beat Pokémon Red working only from screenshots[1].
Pricing: same rates, fewer tokens
Sonnet 5.5 keeps Sonnet 5's list prices, but its improved efficiency means real per-task costs fall[1]:
| Per 1M tokens | Claude Sonnet 5.5 | Claude Opus 5.5 |
|---|---|---|
| Input | $2 | $4 |
| Output | $10 | $20 |
| Cache reads | $0.20 | $0.20 |
| Cache writes | $2.50 | $5 |
API pricing per 1M tokens, September 2026[1].
Anthropic says Sonnet 5.5 generates output more than 30% faster than Sonnet 5, making it the fastest Sonnet model to date. Early testers also observed it batching tool calls together more often, which leads to fewer steps and lower costs on agentic work[1].
Safety and availability
Because Sonnet 5.5's cybersecurity capabilities are comparable to Opus 5's, it's the first Sonnet model to launch with cyber safeguards and fallbacks like those Anthropic built for its most capable models. Its biology safeguards match Sonnet 5's[1].
The smaller Claude Haiku 5.5 — built for high-volume, cost-sensitive applications — was slated to "join the Claude 5.5 family in the coming weeks" at the time of this launch[1].
Bottom line
Verdict: For developers and teams doing everyday coding, debugging, and document work on Claude, Sonnet 5.5 is a clear upgrade over Sonnet 5 — faster, cheaper per task, and dramatically stronger at agentic coding. Keep Opus 5.5 for complex, open-ended work that needs sustained judgment; reach for Sonnet 5.5 for the well-scoped bulk of your workload.
Frequently asked questions
When did Claude Sonnet 5.5 release?
Anthropic announced Claude Sonnet 5.5 on September 28, 2026[1][2].
How much does Claude Sonnet 5.5 cost?
$2 per million input tokens, $10 per million output tokens, and $0.20 per million tokens for cache reads — the same list prices as Sonnet 5, but with up to 30% lower cost per task[1].
Is Sonnet 5.5 faster than Sonnet 5?
Yes — Anthropic reports output generation more than 30% faster than Sonnet 5[1].
How does Sonnet 5.5 compare to Opus 5.5?
Sonnet 5.5 is cheaper and faster, and it's nearly level with Opus 5.5 on some benchmarks, but Opus 5.5 remains stronger at complex, open-ended work requiring sustained judgment[1].
Sources
- Anthropic — "Introducing Claude Sonnet 5.5" — anthropic.com
- Claude Help Center — Release notes (September 2026) — support.claude.com
- Anthropic — "Claude Sonnet 5.5 System Card" — anthropic.com