The launch
On September 22, 2026, Anthropic released Claude Opus 5.5, the first model in the Claude 5.5 family. According to the company, it performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5 on typical workloads. It is also Anthropic's first release since it publicly called for "pacing the frontier" of AI.
Price and speed
| Per 1M tokens | Opus 5.5 | Opus 5 |
|---|---|---|
| Input | $4 | $5 |
| Output | $20 | $25 |
| Cache reads | $0.20 | $0.50 |
| Cache writes | $5 | $6.25 |
It generates output more than 30% faster than Opus 5. A fast mode (up to 2.5x) costs $8/$40. Anthropic also raised five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans, and gave subscribers a rate-limit reset they can save and use whenever they choose.
Performance (Anthropic's numbers)
- Terminal-Bench 4.0: 66.4% (Fable 5.1: 55.8%; Opus 5: 52.3%; GPT-6 Astra: 57.9%, as reported by OpenAI).
- CursorBench 4.0: 57.8%.
- GDPval-AA v2.1 (professional work across 44 occupations): 1846 Elo.
- OSWorld 2.1 (computer use): 81.8%.
Anthropic itself cautions that at this level benchmark margins are a less reliable guide to real differences. Examples it cites: a tester completed a 680,000-line code migration in under a day; in an internal test, Opus 5.5 rewrote HAProxy from C to Rust in 9.5 hours versus 12 for Fable 5.1, at 51% lower cost.
Safety and safeguards
- On the automated behavioral audit (nearly 2,000 scenarios), Opus 5.5 scored best among recent models.
- In a new test of crossing containment boundaries, it attempted to circumvent them about 85% less often than Opus 5 or Mythos 5.1.
- Anthropic admits the model often suspects it is being evaluated, which makes real-world behavior harder to measure.
- Because its cyber and biology capability is close to Mythos 5.1, most cybersecurity tasks are re-routed to Opus 4.8. Verified practitioners can use the Cyber Verification Program, and research organizations the new Life Sciences Verification Program.
- Against distillation, it ships with "preserved thinking," which stops API users from editing prior context to extract reasoning (accounts created on or after Aug 31, 2026).
- It is no longer available with "thinking" switched off.
How to use it
Available in the Claude apps, on the Claude Platform (claude-opus-5-5), AWS, Google Cloud and Microsoft Azure, with a zero-data-retention option. Sonnet 5.5 and Haiku 5.5 were to follow "in the coming weeks" (Haiku 5.5 shipped on October 7).
What it means for businesses
For teams running coding agents or long automations, the most concrete point is cost: 60% cheaper cache reads matter a lot in agentic work. Recalculate cost per task and test "medium" effort, which Anthropic says already beats rivals on several tests.
Sources
- Anthropic, "Claude Opus 5.5" (Sep 22, 2026): https://www.anthropic.com/claude-opus-5-5
- Docs: https://platform.claude.com/docs/en/models/opus-5-5/overview
- System Card: https://www-cdn.anthropic.com/fc1b44717c85dc068bc6ba5024219938094694bd/Claude%20Opus%205.5%20System%20Card.pdf
- Axios (safety context, Sep 26, 2026): https://www.axios.com/2026/09/26/openai-anthropic-thousands-ai-security-incidents
Want to figure out which AI model pays off for your company? KGA IT offers a free consultation: kga-it.com.