Claude Sonnet 5.5: Near-Opus Performance at Half the Price

Six days after Opus 5.5, Anthropic has released Claude Sonnet 5.5. The price is unchanged at $2 input / $10 output per million tokens, the same as Sonnet 5 and the same as OpenAI's GPT-6 Sol. That's half the cost of Opus 5.5.

The biggest jump is in terminal coding. Sonnet 5.5 scores 70.6% on Terminal-Bench 4.0, up from 10.3% for Sonnet 5, and it beats Opus 5.5's 66.4%. On everything else it sits a few points below Opus. On GDPval-AA knowledge work it's 2 points behind (1844 vs 1846), and on OSWorld computer use it scores 80.1% vs 81.8%.

Against GPT-6 Sol at the same price, Sonnet 5.5 leads on knowledge work (1844 vs 1487) and chart reading (61.6% vs 53.6%), while Sol comes out ahead on FrontierCode (49.3% vs 46.2%). Anthropic also says Sonnet 5.5 uses fewer tokens per task, cutting real costs by up to 30%. The company still recommends Opus 5.5 for complex, open-ended work.

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