Every's Vibe Check: Anthropic's Fable 5.1 Is Stronger, Friendlier, and Uses Half the Tokens of Opus 5
every · x · 2026-09-26
Every published a vibe check of Fable 5.1, calling it the strongest coding model they've used while fixing its predecessor's flaws:
- Capability: more powerful than Fable 5, running for days on coding challenges that stumped the older model
- Experience: fast and speaks like a normal person, unlike the original Fable and recent Sonnet 5 / Opus 5
- Cost: uses roughly half the tokens of Opus 5 for similar tasks — Fable-level work within an Opus budget
- Enterprise: new zero-data retention option removes a key barrier for large companies
- Audience: not just for programmers — knowledge workers can genuinely delegate tasks like slide decks and get solid results in an hour or two
Dan Shipper also noted that when he asked Fable 5.1 to analyze a customer survey, it both found the insight and articulated it in one clean sentence — the first time he's seen a model (including earlier Fable and Opus 5) nail precise insight communication.
More from coding & agent
- Dev shows agents now run his entire self-improving eval loop, echoing Anthropic's recursive self-improvement warning — alexcovo_eth · 2026-09-26
- Directory maps 100+ production JEV architectures: LLMs write, JEV decides, code enforces — alexcovo_eth · 2026-09-26
- One-line prompt fix for sycophancy: make the model argue against you before it agrees — victor_explore · 2026-09-26
- Microsoft ships on OpenClaw as industry pivots to always-on agents; OpenAI hired its creator — heyneighbor · 2026-09-26
- Training a 600M model with a moving attention window and memory: what it revealed — KlausCodes · 2026-09-26
- Dev trains a 50MB model in 15 min to replace Gemini Flash at 0.06s latency — newz2000 · 2026-09-26