GPT 6 Astra vs Claude Fable 5.1

Same task, completely different creative instincts.

Neither GPT nor Claude generates video — they orchestrate separate video-gen models. So this isn’t testing GPT-6 vs Fable, it’s testing whatever models sit behind each, with different prompts. Models are probabilistic too — similar prompts, very different outputs.

and we still wont know which is better until we break both on the same prompt

Fable did a better job. Obviously, Astra is new and has more capabilities, but until it’s properly trained, I wouldn't trust it with heavy or critical tasks.

but how do you do this ?

Ad agencies everywhere are crapping themselves!

the naming alone tells you everything, astra sounds like a satellite, fable sounds like it's gonna narrate my exit liquidity

Big brain battle, who wins.

Where you can make this kind of ads? Which program?

Damn astra did a better job by actually showing most scenes with chocolate compared to 5.1 that was more environmental, did you give it the same prompt?

Astra is better

@RoundtableSpace always curious how folks decide which AI to go with. it's like picking a new teammate, gotta find the right fit for your style!

astra winner

Which one wins this round

Astra is clearly better

Astra wins. 🥇

What mcp are using both?

pretty sure the prompt wasn't the same

claude is abosulte shit without mythos

so it uses runaway right to generate it or it has its own model to generate video

For me, it’s hardly surprising, really, since OpenAI’s video model was absolutely amazing! Unfortunately, it is no longer with us - R.I.P. Sora. (P.S.: I know the video wasn’t created by GPT, but the narrative might well have come from Sora 2’s training data.)

Astra made a more realistic ad.

Damn Astra cooked

GPT is better right