Anecdote◆Article
GPT-5.6 Sol is very competent but has not struck me as better than Claude Fable 5 for the complex coding tasks I use.
Based on early access, Simon Willison finds GPT-5.6 Sol competent but not superior to Claude Fable 5 for the complex coding work he does. ✦ AI generated
Simon Willison · Simon Willison's Weblog · 2026-07-09 · original ↗
I've had some early access to GPT-5.6 Sol - it's definitely very competent, though so far it hasn't struck me as better than Fable at the kind of complex coding tasks I've been using with Anthropic's model.
Read full article ↗excerpt · fair-use quotation
- ·Had early access to GPT-5.6 Sol
- ·Says it's 'definitely very competent'
- ·Not better than Claude Fable 5 so far
- ·Judged specifically on complex coding tasks
- ·Hasn't beaten Fable at complex coding work
- ·Fable is Anthropic's model, per Willison
- ·Assessment based on his own usage, not benchmarks
Around this claim
This moment responds to
rebuts → GPT-5.6 Sol is the most practically effective model for real product work, even though Claude Fable may be theoretically smarter.Claire · Lenny's Newsletterrebuts → GPT 5.6 Soul one-shotted a complex CRT/VHS video shader with full control over scan lines, line pitching, bloom, vignetting, and ghosting, a result Scott doesn't believe other models could have achieved.Scott Tolinski · Syntaxrebuts → GPT-5.6 Sol sets a new high score of 53.6 on the Agents' Last Exam benchmark, beating Claude Fable 5 by 13.1 points, and even at medium reasoning it beats Fable 5 by 11.4 points at roughly a quarter of the cost.OpenAI · Simon Willison's Weblogrebuts → Claude Fable gets too locked into its own frameworks and refuses to reconsider, while GPT-5.6 Sol is willing to adapt and fix problems pragmatically.Claire · Lenny's Newsletter