Claude Opus 5.5: Fable-Level Performance at Opus Prices

Anthropic has released Claude Opus 5.5, the first model in its new 5.5 family. The pitch is that it performs at roughly Fable 5.1's level while costing $4 input / $20 output per million tokens, versus $10 / $50 for Fable. It also runs 30%+ faster than Opus 5.

The coding gains are the largest. On Terminal-Bench 4.0 it scores 66.4% against 52.3% for Opus 5 and 57.9% for GPT-6 Astra. In one Anthropic test it audited and fixed a 200K-line codebase in under 3 hours, compared with 20+ hours for Opus 5. It also leads on CursorBench (57.8%) and GDPval-AA knowledge work (1846).

It doesn't win everywhere. GPT-6 Astra still leads on agentic scientific research (64.6% vs 58.7%) and edges it on business workflows (41.4% vs 40%). Sonnet 5.5 and Haiku 5.5 are due in the next few weeks.

Top 7 Perks | Variational Points +12% | Nado Private Invite | Kolo Free Card Issuance | $20 in KAST Points