Hands-on with Claude Sonnet 5.5: half the cost, 30% faster, the go-to execution model
卡尔的AI沃茨 · wechat · 2026-09-29
Anthropic quietly shipped Sonnet 5.5, positioned as the fast, low-cost tier alongside Opus 5.5. The author ran extensive hands-on tests:
- Pricing & performance: same price as Sonnet 5, 50% cheaper than Opus 5.5, 30% faster output, up to 30% cheaper per task; benchmark scores stay close to Opus 5.5 and it even wins on terminal coding.
- Best at: well-scoped daily tasks, bug fixing, docs/PPT/spreadsheets — an execution model.
- Tests: asked both models to build a 30-second site promo purely with code (no video/music/image gen). Opus 5.5 nailed music sync and added unprompted creative touches like liquid physics; Sonnet 5.5 performed like a strong senior frontend dev — clean, complete, but no extra flair. Its 3D skills land around GPT-6sol level; in a writing test its pacing arguably beat Opus 5.5.
- Workflow advice: use Opus for creative direction and aesthetic calls, Sonnet for batch fixes, docs, and deterministic execution — half the cost, no quality loss.
The author also previews OpenAI's DevDay, guessing at a long-running coding agent or an Ultrafast 700 token/s inference tier.
More from coding & agent
- Generic API-catalog MCP server vs one tool per job: a context-bloat experiment — MandeepKaur9407 · 2026-11-16
- Berkeley paper: LLMs answer with stale values even when the update is in context — ombre-ai · 2026-11-16
- Effective HTML: 4.2k-star agent skills turn wireframes into self-contained interactive prototypes — tom_doerr · 2026-11-16
- skills-gate: one gate skill fronts 148 agent skills at just 389 tokens — Distinct-Pie2389 · 2026-11-16
- Open-weight Karpo-Decision-31B tops Decision Index at 73.9, one forward pass per answer — CodeByPoonam · 2026-11-16
- shadcn boosts @uisdk: one-shot UI rebuild that's instantly E2E-testable by AI — shadcn · 2026-11-16