Claude Opus 5.5 is a rare upgrade that costs less than the model it supersedes: every rate — input, output, cache write and cache hit — is lower than Opus 5's. Both keep a 1M-token context at standard pricing.
Claude Opus 5.5 is cheaper on every line: 20% less on input and output, and its cache hit ($0.20) is 60% below Opus 5's $0.50. Unless you have unfinished evals on Opus 5, there is no cost reason to stay.
| Claude Opus 5.5 | Claude Opus 5 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Tier | Flagship | Flagship |
| Input / 1M tokens | $4.00 | $5.00 |
| Output / 1M tokens | $20.00 | $25.00 |
| Cache hit / 1M | $0.20 | $0.50 |
| Cache write / 1M | $5.00 | $6.25 |
| Context window | 1M | 1M |
| Notes | New leading Opus model (Sep 22, 2026); Fable 5.1-level performance at 20% less than Opus 5; included on Pro/Max/Team at no extra cost | New Opus flagship (Jul 24); near-Fable 5 at half the price |
Claude Opus 5.5 is 20% cheaper on input tokens ($4.00 vs $5.00 per 1M). Claude Opus 5.5 is 20% cheaper on output tokens ($20.00 vs $25.00 per 1M).
| Workload | Claude Opus 5.5 | Claude Opus 5 | Difference |
|---|---|---|---|
| Support chatbot 10,000 conversations × (2,000 input + 400 output tokens) | $160 | $200 | Claude Opus 5.5 saves $40.00 |
| Document Q&A 1,000 questions × (30,000 input + 800 output tokens) | $136 | $170 | Claude Opus 5.5 saves $34.00 |
| Coding agent session 4M input tokens (90% cache hits) + 150,000 output tokens | $5.32 | $7.55 | Claude Opus 5.5 saves $2.23 |
Standard rates per 1M tokens from data on this site. The coding-agent row assumes 90% of input is read from the prompt cache (full input price where a model has no cache rate). Cache-write fees, batch discounts and long-context surcharges listed in the notes are not included. Updated November 2, 2026.
you are starting a new project or can re-run your evals — it is cheaper per token, much cheaper on cached input, and has lower fast-mode rates.
you have a production prompt suite validated on Opus 5 and need time to confirm Opus 5.5 behaves the same before migrating.
Anthropic prices Opus 5.5's cache hits at 5% of its base input rate, against the usual 10% on Opus 5. For agentic coding — where most input is cached context re-sent every turn — that means the cached share of the bill falls by more than half, on top of the 20% cut to output. The coding-agent row above captures this.
Anthropic's pricing page says both models bill their full 1M-token context at standard rates, so a 900K-token request pays the same per token as a short one. Both get the 50% Batch API discount. Fast mode, which trades a higher price for faster output, is available on Opus 5.5 at lower rates than on Opus 5, where it is still a research preview.
Because pricing is strictly lower, the only real cost of upgrading is engineering time: re-run your regression prompts, check output length (a model that writes more can erase a per-token saving), and verify tool-call formats. Opus 5.5 is also included on Claude Pro, Max and Team plans at no extra cost, so teams can compare both in the app first.
Yes. Opus 5.5 is $4.00 input and $20.00 output per 1M tokens versus $5.00 and $25.00, with cache writes at $5.00 vs $6.25.
No. Anthropic bills the full 1M-token context at standard pricing for both.
Yes, Anthropic's Batch API halves input and output prices on both models.
How we choose: prices come from each vendor's official pricing page and are re-checked every week. We take no affiliate fees, so rankings aren't for sale.