benchrHead-to-head: how this rate card stacks against OpenAI's daily driver is covered in Sonnet 4.6 vs GPT-5, and against the open-weight challenger in DeepSeek V4-Pro vs Sonnet 4.6.
Pricing breakdown
| Tier | Rate / 1M tokens |
|---|---|
| Standard input | $3.00 |
| Standard output | $15.00 |
| Cached input | $0.30 |
| Context window | 200,000 tokens |
| Max output | 64,000 tokens |
Why Sonnet is the default, not Opus
Most production LLM traffic does not require frontier-level reasoning. Sonnet 4.6 handles conversational tasks, document summarization, code review, moderate-complexity code generation, and classification at 79.6% SWE-bench — strong enough for the vast majority of real-world use cases. The 9-point gap to Opus 4.8 matters specifically when tasks require multi-step reasoning over complex codebases, where errors compound and the quality ceiling is actually reached.
Anthropic positions Sonnet as the production default for this reason. The cost economics favor defaulting to Sonnet and upgrading only where evaluation data shows Opus delivering better task outcomes — not as a blanket upgrade.
The 64K max output advantage
Claude Sonnet 4.6 supports 64,000 output tokens per response — twice the Opus limit of 32K. For output-heavy workloads — generating full documentation, producing large code files, writing long-form reports — Sonnet 4.6 is better suited than Opus regardless of reasoning depth. A 64K output at Sonnet pricing ($15/1M) costs $0.96 per call; the same at Opus 4.8 pricing ($25/1M) would cost $1.60 and would require splitting into two calls.
Caching and effective rates
Claude Sonnet 4.6 cached input costs $0.30/1M — 90% off the $3 standard rate. For a typical agent with a 50K-token system prompt running 1,000 calls per day, caching saves approximately $135/day versus uncached at $3 per million. With high cache hit rates, the effective input cost drops to roughly $0.57 per million — competitive with Haiku uncached pricing for input-heavy workloads.
Cost scenarios
At 20M input + 5M output per month: $60 + $75 = $135/month. Opus 4.8 at the same volume: $100 + $125 = $225/month — $90 more for the Opus premium. With 90% cache hit rate on Sonnet 4.6: approximately $11.40 + $75 = $86.40/month. For a startup processing customer queries, this is a practical monthly budget difference that compounds at scale.
Use-case fit
Best for: Production chat and assistant applications; code review and moderate-complexity generation; summarization of long documents (up to 200K context); pipelines where 64K output length is an advantage; cost-sensitive deployments where frontier performance isn't required.
Skip if: Your tasks require frontier-level multi-step coding (use Opus 4.8); or you need maximum throughput on simple tasks at minimal cost (use Haiku 4.5 at $1/1M).
Decision checklist
Before using Opus 4.8 instead: run a 50-sample eval of your hardest task category on both Sonnet 4.6 and Opus 4.8. If you can't detect a quality difference, stay on Sonnet — the 40% cost reduction compounds significantly at production volume.
Before using Haiku 4.5 instead: confirm your tasks don't require sustained reasoning. Haiku's 73.3% SWE-bench score is a 6-point drop from Sonnet's 79.6%, which is measurable in coding and analysis tasks even if invisible in simple classification.