Anthropic's new leading Opus model — near-Fable 5.1 performance at 40% less than Opus 5. Both stay listed since 5.5 launched at a different price.
Claude Opus 5.5, released September 22, 2026, is Anthropic's new leading Opus-tier model, priced at $4 per 1M input tokens and $20 per 1M output tokens — 40% cheaper than Opus 5's $5/$25, while Anthropic says it performs at roughly Fable 5.1 level on most work. Because it launched at a different price than Opus 5 rather than matching it, both models stay listed side by side on this site: Opus 5.5 is the new default recommendation, but Opus 5 remains callable on the API.
Anthropic reports Opus 5.5 also generates output over 30% faster than Opus 5. In the consumer apps, it appeared in the model menu on Claude Pro, Max, Team and seat-based Enterprise plans on day one — Pro subscribers get it without paying extra — and Anthropic raised the rolling 5-hour usage limits on those plans alongside the launch.
Cache economics improved too: a cache hit on Opus 5.5 costs 5% of the base input price ($0.20 per 1M), versus the standard 10% multiplier most other Claude models use. A 5-minute cache write is $5 per 1M and a 1-hour cache write is $8 per 1M.
Opus 5.5 targets the same workloads as Opus 5 — agentic coding, long-document analysis, and enterprise deliverables — at a lower cost per task. A fast mode (research preview, Claude API only) runs at premium pricing of $8/$40 per 1M for faster output. Like the rest of the Claude 4.6+/5.x line, it ships the full 1M-token context window at standard pricing, so a 900K-token request costs the same per token as a 9K-token one.
As with Opus 5, Anthropic does not offer image, video, or audio generation on this model — Claude stays text-and-analysis only.
| Model | Input / 1M | Output / 1M | Context |
|---|---|---|---|
| Claude Fable 5.1 | $10 | $50 | 1M |
| Claude Opus 5.5 | $4 | $20 | 1M |
| Claude Opus 5 | $5 | $25 | 1M |
| Claude Sonnet 5.5 | $2 | $10 | 1M |
| Claude Haiku 4.5 | $1 | $5 | 200K |
| GPT-5.5 | $5 | $30 | 1M |
Opus 5.5 undercuts Opus 5 by 20% on both input and output while claiming most of Fable 5.1's quality — a sharper price/performance point in the Opus tier. Because the launch price differs from Opus 5's, both stay on the API rather than one fully replacing the other. Within the family, Sonnet 5 is still the value pick at under half of Opus 5.5's price, and Fable 5.1 stays on top for the very hardest long-horizon work.