Comparison
Claude Sonnet vs Opus: speed and budget trade-offs
Both Claude profiles serve the Anthropic-compatible endpoint. Sonnet is the workhorse; Opus is the escalation tier for the hardest reviews, at a higher price.
Updated
| Criterion | Claude Sonnet | Claude Opus |
|---|---|---|
| Price tier | Lower per token | Higher per token |
| Best for | Everyday edits, drafts, agent loops | Deep review, hard reasoning, long instructions |
| Protocol | Anthropic Messages or OpenAI shape | Anthropic Messages or OpenAI shape |
| Cached input | Has a cache price | Has a cache price |
| Billing | Same USD balance, same key | Same USD balance, same key |
Use Sonnet by default, Opus for the hard 10%
Most edits, drafts and agent steps do not need the flagship price. Sonnet covers that work cheaply.
Reach for Opus when a diff review, a long instruction or a subtle bug resists the cheaper model.
Long context is expensive on either
Long prompts and large outputs dominate Claude workloads. Check the actual token counts in the journal, not the headline per-1M number by itself.
Cache the repeated prefix when the provider reports it; the cache rate is cheaper on both profiles.
What we do not claim
We do not publish latency or benchmark numbers we have not measured ourselves. Treat any fixed superiority claim as marketing.
Run your own prompt on both profiles, compare the answer and the billed amount, and let the journal decide.
Published model prices
USD per 1M tokens. The numbers come from the public catalog on every page rebuild — this is the sale price, not an estimate.
| Model | Input | Output | Cache |
|---|---|---|---|
| Claude Opus 5.5 | $2.70 | $8.00 | $0.14 |
| DeepSeek Flash | $0.50 | $2.00 | $0.01 |
| Claude Sonnet 5.5 | $1.33 | $6.67 | $0.07 |
| DeepSeek V4 Pro | $2.20 | $6.60 | $0.07 |
Top-up from $10, balance in USD. 1 USD = 1,000,000 UT. No subscription and no auto-charge.
Questions
Which Claude profile is cheaper?
Sonnet is priced lower per token in the catalog. The exact difference is in the live table above.
Can I use both with Claude Code?
Yes. Set ANTHROPIC_BASE_URL to https://llmcode.io, choose the model per context.
Do Sonnet and Opus share a cache price?
Each has its own cache price, listed in /models.
Which should I default to?
Sonnet for routine work, Opus for the hard 10% after the cheaper model stalls.