benchrPricing breakdown
| Tier | Rate / 1M tokens |
|---|---|
| Standard input | $5.00 |
| Standard output | $25.00 |
| Cached input | $0.50 |
| Context window | 1,000,000 tokens |
| Max output | 32,000 tokens |
Opus 4.7 vs Opus 4.8: the 1-point question
Anthropic reports SWE-bench Verified at 87.6% for Opus 4.7 and 88.6% for Opus 4.8. A one-point benchmark delta does not mean one additional production issue resolved per 100 tasks, because repository mix, prompts, tools, scaffolding, and pass criteria differ.
For an existing 4.7 deployment, compare migration effort with a measured change in task success, review time, latency, and regressions. For a new project, 4.8 is a candidate because its listed token rate is the same, but it still requires workload testing.
Opus vs Sonnet: the real cost decision
Claude Sonnet 4.6 is listed at $3/1M input and $15/1M output — 40% below Opus on both rates. Anthropic reports 79.6% on SWE-bench for Sonnet and 87.6% for Opus 4.7. That eight-point provider benchmark gap does not establish which tier is more economical for refactoring, architecture review, chat, summarization, or classification.
Run a blind evaluation on the same tasks. If Opus does not produce a measurable improvement in accepted outputs or review cost, Sonnet remains the lower-priced candidate.
Caching and the effective rate
Cached input for Claude Opus 4.7 costs $0.50/1M — 90% off standard. For agents with a large fixed system prompt (say 80K tokens of tool definitions and instructions), a 90% cache hit rate brings the effective input cost to roughly $0.95 per million tokens — comparable to uncached Sonnet 4.6 at $3 per million. Output doesn't cache, so output-heavy workloads still pay the full Opus premium on the output side.
Cost scenarios
At medium volume — 5M input tokens and 1.5M output per month — Opus 4.7 costs $25 + $37.50 = $62.50/month. Sonnet 4.6 at the same volume: $15 + $22.50 = $37.50/month. The Opus premium is $25/month at this scale. At 90% cache hit on input: approximately $4.75 + $37.50 = $42.25/month. That puts cached Opus within 13% of uncached Sonnet for systems with stable, reusable context.
Use-case fit
Candidate for: Existing deployments already pinned to 4.7 and workloads such as multi-file coding, architecture review, long-document analysis, or agent loops where a matched evaluation supports the tier.
Compare alternatives if: You are starting a new project or Sonnet passes the same acceptance tests. Include Opus 4.8 because its listed token rates match 4.7.
Decision checklist
Run your task sample on both Claude Sonnet 4.6 and Claude Opus 4.7 in a blind evaluation before committing to Opus pricing. If you cannot detect a useful difference, Sonnet is the lower-priced candidate.
For Opus 4.7 versus 4.8, do not convert the one-point SWE-bench delta into an assumed one-point production gain. Use versioned regression cases and your own pass criteria.