GPT-5.6 Sol vs. Claude Fable: Why OpenAI’s new model crushes my benchmark
A creator conducts a detailed benchmark comparing OpenAI's new GPT-5.6 models (Sol, Terra, Luna) against Claude Fable across real-world tasks like PRD writing, prototyping, and coding. GPT-5.6 Sol emerges as the superior choice for practical product development, offering better design uniqueness, clearer communication, and more effective collaboration despite Fable's theoretical intelligence.
Summary
The creator presents a comprehensive evaluation of OpenAI's newly released GPT-5.6 family of models compared to Anthropic's Claude Fable. The GPT-5.6 lineup consists of three tiers: Sol (frontier model for advanced tasks), Terra (balanced everyday model), and Luna (affordable high-volume option). Sol is priced more competitively than Fable at $5 per million input tokens versus $10.
The evaluation methodology uses a custom 'How I AI Vibe Review Benchmark' that tests models across PRD generation, prototyping, wireframing, code debugging, and agentic voice interactions. The creator weights her own subjective assessment 70% and LLM-based judge scores 30%, revealing a strong preference for Sol's outputs. Specific strengths noted include superior design aesthetics (avoiding generic 'slop'), functional prototypes that work on first attempt, and cleaner visual hierarchy with semantic color usage.
The creator emphasizes that Fable excels at theoretical intelligence and technical precision but struggles with practical effectiveness and human communication. She describes Fable as 'by agents for agents' communication that feels pedantic and difficult to collaborate with, whereas Sol demonstrates better understanding of end-user goals and willingness to reconsider constraints. Key differences manifested in product development projects where Fable's overly-hardened architecture broke implementations, while Sol could pivot and find working solutions.
Additional use cases where Sol particularly excels include video editing and browser automation. The creator demonstrates using Sol to clip social media videos from conference talks and to automate LinkedIn message management through browser use commands. Overall, the assessment positions Sol as practically superior for shipping products and effective collaboration, despite acknowledging Fable's strengths in specific technical domains.
About this episode
<p>GPT-5.6 Sol is back, and I ran it through my full How I AI vibe benchmark against GPT-5.6 Terra, Luna, Claude Fable 5, and Sonnet 5 across five categories: PRDs, prototypes, wireframes, debugging, and agentic voice. Sol won by a meaningful margin on my Claire Weighted Index (70% my taste, 30% Terminal Bench 2.1), and I also tested two use cases I can't stop thinking about: building a gamified homework tracking app for my kids in one shot with Codex, and browser automation with Chrome that burned through 500 LinkedIn replies while I did literally nothing.</p><p><br /></p><p><strong>What you’ll learn:</strong></p><ol><li>How I scored five AI models (including GPT 5.6 Sol, Fable 5, and Sonnet 5) using my “Claire Weighted Index” benchmark across PRDs, prototypes, code, and agentic voice</li><li>The difference between GPT-5.6 Sol (Terra) and Sol for PRD writing</li><li>How Fable’s precision and pedantry made it harder to collaborate with, and the exact moment Sol broke through where Fable got stuck</li><li>Why Sonnet 5 is still my go-to for agentic voice in OpenClaw, even after this whole benchmark</li><li>How I used GPT-5.6 Sol in Codex to build a fully gamified homework tracking app for my kids in one shot</li><li>The video editing use case that saved me hours clipping a talk I gave at Cursor’s event</li><li>How to use Codex plus GPT-5.6 and Chrome for browser automation, and why this is my single most-loved use case right now</li></ol><p>—</p><p><strong>In this episode, I cover:</strong></p><p>(00:00) Intro</p><p>(01:10) The three GPT-5.6 models: Sol, Terra, Luna</p><p>(02:17) Pricing: Sol vs. Fable API costs</p><p>(03:24) The How I AI benchmark</p><p>(05:03) Claire-weighted Index results</p><p>(07:00) Per-task winners: prototypes, PRDs, agentic voice</p><p>(11:59) What Claire actually rewards</p><p>(13:20) Full-fidelity prototype side-by-sides (Sol vs. Fable)</p><p>(17:45) Wireframes</p><p>(18:19) Agentic voice</p><p>(19:15) Where Sol is better than other models</p><p>(23:56) Gamified kids’ homework app, built in one shot</p><p>(28:02) Fable’s pedantry problem and how Sol broke through it</p><p>(31:49) Two bonus use cases: video editing and browser use</p><p>(35:08) Final summary and model recommendations</p><p>—</p><p><strong>Tools referenced:</strong></p><p>• GPT 5.6 (Sol, Terra, Luna): <a href="https://help.openai.com/en/articles/20001325-a-preview-of-gpt-56-sol-terra-and-luna" rel="ugc noopener noreferrer" target="_blank">https://help.openai.com/en/articles/20001325-a-preview-of-gpt-56-sol-terra-and-luna</a></p><p>• Codex: <a href="https://openai.com/codex" rel="ugc noopener noreferrer" target="_blank">https://openai.com/codex</a></p><p>• ChatPRD: <a href="https://www.chatprd.ai/" rel="ugc noopener noreferrer" target="_blank">https://www.chatprd.ai/</a></p><p>• CapCut: <a href="https://www.capcut.com/" rel="ugc noopener noreferrer" target="_blank">https://www.capcut.com/</a></p><p>• Math Academy: <a href="https://www.mathacademy.com/" rel="ugc noopener noreferrer" target="_blank">https://www.mathacademy.com/</a></p><p>—</p><p><strong>Other references:</strong></p><p>• Cursor event where Claire spoke on the future of PM: <a href="https://www.youtube.com/watch?v=4CAFK-rc26A" rel="ugc noopener noreferrer" target="_blank">https://www.youtube.com/watch?v=4CAFK-rc26A</a></p><p>• ChatPRD blog (where benchmark outputs will be published): <a href="https://www.chatprd.ai/" rel="ugc noopener noreferrer" target="_blank">https://www.chatprd.ai/</a></p><p>—</p><p><strong>Where to find Claire Vo:</strong></p><p>ChatPRD: <a href="https://www.chatprd.ai/" rel="ugc noopener noreferrer" target="_blank">https://www.chatprd.ai/</a></p><p>Website: <a href="https://clairevo.com/" rel="ugc noopener noreferrer" target="_blank">https://clairevo.com/</a></p><p>LinkedIn: <a href="https://www.linkedin.com/in/clairevo/" rel="ugc noopener noreferrer" target="_blank">https://www.linkedin.com/in/clairevo/</a></p><p>X: <a href="https://x.com/clairevo" rel="ugc noopener noreferrer" target="_blank">https://x.com/clairevo</a></p><p>—</p><p>Production and marketing by <a href="https://penname.co/" rel="ugc noopener noreferrer" target="_blank">https://penname.co/</a>. For inquiries about sponsoring the podcast, email [email protected].</p>
Key Insights
- The creator developed a weighted evaluation system treating her subjective assessment as 70% of the score and LLM judge scores as 30%, reflecting a philosophy that personal taste matters more than automated benchmarks
- Sol demonstrates superior design originality by avoiding generic 'slop' patterns, using semantic color assignment and functional visual hierarchy, while Fable and Claude models tend toward predictable design aesthetics with characteristic features like beige backgrounds and burnt orange accents
- Fable's primary weakness in practical product development stems not from intelligence but from pedantic precision and hardened architectural decisions that prevent flexibility, whereas Sol shows willingness to reconsider constraints to achieve working solutions
- The creator argues that product building requires intuition, creativity, and boldness rather than technical precision, making Sol's practical effectiveness more valuable than Fable's theoretical rigor for shipping products
- Sol excels at zero-to-one prototypes by generating complex, functional applications in single attempts, such as a fully gamified homework tracking system with parent controls and reward systems that required minimal iteration
- Communication style significantly impacts collaboration effectiveness, with Sol using clear, straightforward writing while Fable employs em-dashes and pedantic phrasing that the creator describes as difficult for humans to parse and collaborate with
- The creator discovered Sol's consistent use of forest green in designs appears to stem from system prompts, representing a recognizable model 'tell' that becomes frustrating through repetition across multiple outputs
- Browser automation combined with GPT-5.6 enables substantial time savings on tasks like social media management, where the creator successfully automated LinkedIn message replies across 500+ messages with quality filtering for high-value connections
Topics
Transcript
I have been very, very, very sad the last week because for the last week, I have not had access to my true favorite top of the line model, GPT-5-6. But guess what, babes? It is back and I am here to walk you through GPT-5-6 Sol, GPT-5-6 Luna, GPT-5-6 Terra. I'm going to tell you what are these models, how have I been using them, why are they my heart's favorite, and is Fable better than all of them or not? I have been testing this model for a couple weeks. There was a few days there where we didn't have access, and I found myself desperate to get this workhorse model back. Now, we're not just relying on…
Full transcript available for MurmurCast members
Sign Up to AccessMore from How I AI
This solo builder runs 24/7 local AI on his own hardware | Alex Finn
Alex Finn, a solo builder, discusses his setup of multiple local AI machines (Mac Studios, DGX Spark, RTX 5090) running 24/7 to power autonomous workflows for software development, security scanning, and market research. He explains why local models unlock unlimited AI usage compared to cloud APIs, and demonstrates his 'software factory' with automated build and review loops using Claude Code.
What a harness is and how to build one with Claude Agent SDK
A harness is custom code built around an AI agent to make it more effective for specific workflows. The speaker demonstrates building a Sentry bug-fixing harness using Claude Agent SDK, explaining how structured constraints, opinionated tooling, and custom prompts enable AI agents to solve complex tasks more reliably than general-purpose coding tools.
How I run autonomous coding agents from my phone with OpenAI Symphony + Linear | Alessio Fanelli (Kernel Labs)
Alessio Fanelli demonstrates how to run autonomous coding agents from a phone using OpenAI's Symphony framework integrated with Linear, showing practical applications from Pokemon card arbitrage to small business automation. He emphasizes moving from agent prompting to agent management through cloud-based orchestration and shares how AI enables efficient scaling of traditionally manual, inefficient business processes.
Sonnet 5 review: I ran 64 generations to find out if it's worth it
The creator develops the 'How I AI Bench,' a custom evaluation framework combining human vibe checks with LLM judging to assess AI models on practical tasks like PRD writing, prototyping, and agentic coding. Testing five frontier models including Claude Sonnet 5, the results reveal significant disagreements between automated scoring and human preference, with Sonnet 4.6 and Gemini 3 Pro ranking highest despite Sonnet 5 being the newest release.