VOD OTT Platform

GPT-5.6 Sol vs Fable 5: I Wasn't Expecting This

2026 13m Subscription

My AI courses: https://edwarddonner.com/curriculum OpenAI dropped GPT 5.6 Sol yesterday. The internet is scrambling to answer the only question that matters. Is it stronger than Claude Fable 5? OpenAI’s splashy announcement had the charts to convince us: Sol is superior. In almost every dimension. Smarter and cheaper. Now, Sol was soundly beaten by Fable in 1 benchmark: SWE-Bench Pro. But the day before, OpenAI declared that SWE-Bench Pro has fundamental issues, is flawed, and no longer meaningful. So that’s that! Naturally, I wanted to see for myself. And I had a great real-world test to hand: the “build a simple Salesforce in 1 shot” challenge that I put to Fable, GLM 5.2 and DeepSeek V4 last week. In this video, I get Sol to build a CRM. We compare it with Fable. It’s marginally cheaper. But is it better? We also look at benchmarks from independent labs. Although there’s probably some benchmaxxing going on. If only there were an indisputable, highly scientific way to judge which model is better. Fortunately, there is. My Connect Four arena has been waiting for such a moment. I can only afford one game. Winner takes all. Sol versus Fable. __ OpenAI's announcement: https://openai.com/index/gpt-5-6/ Their GPT 5.6 developer guide: https://developers.openai.com/api/docs/guides/latest-model Artificial Analysis: https://artificialanalysis.ai/ Subscribe for 2 times a week AI technical breakdowns, and comment which model architecture you want me to test against next. -- 00:00 Is GPT-5.6 Sol Better Than Claude Fable 5? 00:50 What Do OpenAI's GPT-5.6 Benchmarks Actually Show? 01:17 Do Independent Benchmarks Agree With OpenAI? 02:15 Where Does GPT-5.6 Sol Rank on Artificial Analysis? 02:53 GPT-5.6 Sol vs Claude Fable 5: Intelligence vs Cost 03:36 Can GPT-5.6 Build a Production CRM? 05:27 GPT-5.6 Sol CRM Review: UI, Features & Code Quality 07:25 GPT-5.6 Sol vs Claude Fable 5 CRM Comparison 08:29 Is GPT-5.6 Better Value Than Fable 5? 09:17 What's the Best AI Benchmark? Time for Connect 4 09:38 GPT-5.6 Sol vs Claude Fable 5 Connect 4 Showdown 11:55 Final Verdict: Should You Choose GPT-5.6 or Claude Fable 5? -- This is the list of the AI models, technologies, benchmarks, and tools mentioned in the video: GPT-5.6 Sol Claude Fable 5 OpenAI Anthropic Artificial Analysis

Sign in to watch 🔒 Subscription required
Cast
Edward Donner
In
Chatgpt
Views
0