After a day of real use: Sonnet 5.5 is not just Opus 5.5 at half price
alexcovo_eth · x · 2026-09-29
Developer @kunchenguid shares hands-on observations after a day of using Sonnet 5.5 and Opus 5.5:
- Same lineage: The two models behave remarkably alike in style and personality; he speculates both were distilled from a shared teacher model, possibly "Mythos 5.5" or even 6.
- Not a drop-in substitute: "Sonnet = Opus at 50% off" only holds for problems that don't require much wisdom. When asked to plan a complex feature with unclear direction, the gap shows: Opus reasons strategically about "what's the real goal here," while Sonnet thinks tactically about "how do I get this done."
- Workflow: He uses Sonnet 5.5 heavily as his coding first-mate and even prefers it over Opus — faster, cheaper, and smart enough to follow his intent.
Bottom line: the gap appears on ambiguous problems requiring judgment; for everyday coding, Sonnet offers better value.
More from coding & agent
- eqsl-mcp: An MCP Server That Brings Natural Language to Amateur Radio QSL Logs — modelcontextprotocol · 2026-09-29
- After 75 interviews in 4 countries, one year proved 'AI can't code' wrong — realmeetjames · 2026-09-29
- 400 LLM Agents Living in an MMO Server: Lessons on Stale Perception, Fire-and-Forget Actions, and Load Shedding — kristiantalley679 · 2026-09-29
- MemoryDesk: A Support Agent That Remembers Which Fixes Failed Across Sessions — jagarlapudi_radhakri · 2026-09-29
- Agent turns two NeurIPS papers into a 3-minute fully code-rendered launch film — liuziwei7 · 2026-09-29
- Giving a social media agent persistent memory with an experiential layer — Premsai04_ · 2026-09-29