GPT 5.6 Sol Claims Opus-Class Performance Cheaper Than Opus 4.8

GPT 5.6 sol and Fable are NOT the same

> GPT 5.6 sol - Opus class model but faster and cheaper > Fable - bigger model, ideal for very complex coding tasks

Both models are great in their class

On the eve of it's 250th birthday, the US leads in AI 🚀

@bindureddy OpenAI didn't build Sol to kill Fable. They built it to make everyone question why they're still paying Opus 4.8 prices

@IrmaMRo Well who cares about anything as long as the output is just as good, right?

Ok then, Kimi 2.7 is just as good for 99% of all tasks

@bindureddy Who cares as long as the output is just as good? Even if it's almost as good but faster, cheaper and available, it will be a banger

@bindureddy Actually the new bar is GPT5.6 Class Model, all models are now worse.

@bindureddy nonsense claim, its not even been widely tested yet

@bindureddy The gap between models is shrinking faster than the marketing claims.

@bindureddy Absolutely, it's not as powerful as Fable 5

@AdityaKTech exactly

@bindureddy Not really. Not for my work. For academic research, so far only chat GPT and Claude are actually usable.

@bindureddy everyone's calling everything "opus class" this year, at some point that phrase stops meaning anything

@bindureddy Hmmm in my own experience GPT-5.5 is already Opus class.

@bindureddy

@bindureddy finally you stop to talk nonsense, I agree 100% with this post.

@bindureddy opus 4.8 itself was already a tier above most people using it daily

this feels like a flex on a flex

@bindureddy Good to see the distinct use cases for GPT 5.6 sol and Fable, and that US innovation continues to push AI forward.

@bindureddy it's not "Opus class"... 5.5 > Opus already

Right that they're different models - but the tiers are flipped. Sol is OpenAI's flagship at $5/$30 per 1M, same price as GPT-5.5, and it's sitting in a ~20-org government-coordinated preview, not in ChatGPT. Terra is the 2x-cheaper one. Which of them did you actually get access to?

@bindureddy That classification actually mismatches the benchmark data released in the recent system cards. Sol is OpenAI's absolute flagship tier designed to compete at the very top of the reasoning bracket rather than acting as a budget alternative

@bindureddy @abacusai the US leads in regulations and China leads in research and open source technologies