Side by side pricing and context for o3 and Claude Opus 4.6. Rates are curated catalog values with last verified dates on each model page.
o3
OpenAI reasoning model for harder problems where extra thinking tokens are worth the bill. Watch output cost: reasoning tokens are billed at the output rate.
Input $2.00 / Output $8.00 per 1M
Context 200K · Exact (o200k_base)
Claude Opus 4.6
Prior Opus-class Claude for deep work when you need Opus capability at published Opus rates. Prefer for quality-critical paths, not bulk classification.
Input $5.00 / Output $25.00 per 1M
Context 1M · Approx (Anthropic family)
Quick take
- o3 is cheaper on input in this catalog snapshot.
- o3 is cheaper on output in this catalog snapshot.
- Both last verified on their model pages (2026-08-02 / 2026-08-02).
Rate table
| Metric | o3 | Claude Opus 4.6 |
|---|---|---|
| Input / 1M | $2.00 | $5.00 |
| Output / 1M | $8.00 | $25.00 |
| Cached input / 1M | $0.500 | $0.500 |
| Context | 200K | 1M |
| Tokenizer | Exact (o200k_base) | Approx (Anthropic family) |