Claude Sonnet 5.5 jumps from 10.3% to 70.6% on Terminal-Bench 4.0, runs 30% faster at same price
claudeai · x · 2026-09-29
Anthropic releases Claude Sonnet 5.5, a faster, cheaper complement to Opus 5.5 tuned for well-scoped everyday work: bug fixing, polished docs/slides/spreadsheets, and strong design sense. Haiku 5.5 follows in coming weeks.
Key facts:
- Performance: scores 70.6% on Terminal-Bench 4.0 (agentic coding) vs Sonnet 5's 10.3%; only 2 points below Opus 5.5 on GDPval-AA; first Sonnet to beat Pokémon Red from screenshots alone, with strong long-horizon and image understanding.
- Pricing: unchanged from Sonnet 5 ($2/M input, $10/M output, $0.20/M cache reads), but it needs far fewer tokens — up to 30% cheaper per task in testing.
- Speed: 30%+ faster generation, the fastest Sonnet yet.
- Safety: matches or improves Sonnet 5 on most alignment/honesty measures in automated behavioral audits; first Sonnet with cyber safeguards and fallbacks on par with flagship models, without impacting routine dev work.
Early testers describe it as a better collaboration partner with noticeably clearer writing.
Related event: Anthropic Launches Claude Sonnet 5.5: Over 30% Faster, Up to 30% Cheaper(58 posts)→
More from Models
- Jason Calacanis threatens to ditch $20k/yr Claude as $40/mo Grok Bot outperforms it — chrisfleck · 2026-11-13
- Dan Shipper's weekend plan: let Claude Fable 5.2 build a custom Supernote OS — danshipper · 2026-11-13
- Alibaba teases Qwen4 with stronger coding and agentic capabilities at Nov 17 Hong Kong summit — teortaxesTex · 2026-11-13
- Baidu opens ERNIE 5.5 Flash Preview via OpenRouter — Baidu_Inc · 2026-11-13
- Hybrid Decision Model learns when to reason: 72.5% lower latency at higher accuracy — hbouammar · 2026-11-13
- AI×Bio benchmarks called out as LLM-generated "benchslop" that may be impairing bio-reasoning — anshulkundaje · 2026-11-13