Claude Opus 5.5 and Sonnet 5.5: What Changed
Short answer
Anthropic announced Claude Opus 5.5 on 22 September 2026 at $4 input / $20 output per million tokens, saying it performs at the level of Fable 5.1 on most work and costs 40% less to run than Opus 5. Sonnet 5.5 followed on 28 September at $2 / $10, with Anthropic claiming 30%+ faster output than Sonnet 5.
For most writing, analysis and coding, Sonnet 5.5 is the cost-sensible default; use Opus 5.5 for long, open-ended work. Consumer plan prices are unchanged: Pro is $20 a month.
Key facts
| Opus 5.5 | Sonnet 5.5 | Evidence | |
|---|---|---|---|
| Announced | 22 September 2026 | 28 September 2026 | Verified |
| API price per 1M tokens (input / output) | $4 / $20 | $2 / $10 | Verified |
| Cache reads per 1M tokens | $0.20 | $0.20 | Verified |
| Available on | Pro, Max, Team, Enterprise; AWS, Google Cloud, Azure | AWS, Google Cloud, Azure and Claude plans | Verified |
| Speed vs predecessor | More than 30% faster than Opus 5 | 30%+ faster than Sonnet 5 | Claimed by provider |
| Terminal-Bench 4.0 | 66.4% | 70.6% | Claimed by provider |
| OSWorld 2.1 (partial) | 81.8% | 80.1% | Claimed by provider |
| Context window | Up to 1M tokens across Claude models (per-model limit not stated) | Same | Verified (range only) |
What it means for buyers
Evidence: VerifiedAPI users get a price cut: Opus 5.5 at $4 / $20 per million tokens and Sonnet 5.5 at $2 / $10. Anthropic says Sonnet 5.5 scores nearly level with Opus 5.5 on GDPval-AA v2.1, so test Sonnet first before paying twice as much for Opus. Subscribers on Pro, Max and Team get the new models without a price change; Anthropic also said it was increasing five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans with the Opus 5.5 release.
When Sonnet 5.5 is enough
Evidence: Editorial assessment- Enough: drafting, summarising, code review, spreadsheets and documents, high-volume API work.
- Consider Opus 5.5: long multi-step projects, ambiguous requirements, long agent runs.
- Benchmarks are vendor-reported and mixed across tasks; Sonnet 5.5 leads Opus 5.5 on Terminal-Bench 4.0 in Anthropic's own figures. Test on your own work.
Related releases
Evidence: VerifiedFable 5.1 and Mythos 5.1 were announced on 1 September 2026 and sit above Opus; Fable 5.1 costs $10 / $50 per million tokens and Mythos 5.1 is restricted to vetted US organisations. Earlier, on 12 June 2026, Anthropic suspended access to Fable 5 and Mythos 5 after a US government directive.
Pages we updated
Common questions
How much does Claude Opus 5.5 cost?
$4 per million input tokens and $20 per million output tokens on the API, according to Anthropic's announcement and pricing page. In the Claude apps it is included in Pro, Max, Team and Enterprise plans.
Is Sonnet 5.5 better than Opus 5.5?
Not across the board. Anthropic says Sonnet 5.5 scores nearly level with Opus 5.5 on GDPval-AA and higher on Terminal-Bench 4.0, at half the API price. Opus 5.5 remains Anthropic's choice for complex open-ended work.
Did Claude Pro get more expensive?
No. Claude Pro remains $20 a month ($17 billed annually), checked on 5 October 2026.
Sources
- Anthropic: Introducing Claude Opus 5.5 · checked 5 October 2026
- Anthropic: Introducing Claude Sonnet 5.5 · checked 5 October 2026
- Claude pricing · checked 5 October 2026
- Anthropic newsroom and Fable 5.1 / Mythos 5.1 announcement · checked 5 October 2026
- Anthropic statement on Fable 5 and Mythos 5 access · checked 5 October 2026