A.I.
Anthropic
Claude frontier lineup
| Model | Status | AA score | Price | Released |
|---|---|---|---|---|
| Claude Opus 5.5 First Claude 5.5 model. 1M · 128K · $4/$20 · adaptive · claude-opus-5-5 | Current | 58 (max effort) | $4 in, $20 out | 22 Sep 2026 |
| Claude Sonnet 5.5 Second Claude 5.5 model. 1M · 128K · $2/$10 · adaptive · claude-sonnet-5-5 | Current | 56 (max effort) | $2 in, $10 out | 28 Sep 2026 |
| Claude Haiku 5.5 Third Claude 5.5 model. 1M · 128K · from $0.10/$0.50 · adaptive · claude-haiku-5-5 | Current | 43 (max effort) | $0.10 in, $0.50 out | 7 Oct 2026 |
| Claude Fable 5.1 Most capable widely released. 1M ctx · 128K out · $10/$50 · adaptive always onMythos 5.1 = same, Glasswing only. | Current | 53 (max effort) | $10 in, $50 out | Not on file |
3 older models from this lab are in the archive.
AA score is the Artificial Analysis Intelligence Index, one independent score built from the same tests on every model. Prices are US dollars per million tokens (word pieces) in and out unless noted. "Not on file" means this site has no checked value yet.
Products and features
Apps and agents from Anthropic. They run on Claude but are not API models.
- Claude Duet: always-on personal agent research preview.