Artificial Analysis 1 year ago today.
Crazy to think GPT OSS 120B a model with only 5b active parameters was the 8th best model in the world.
A truly extraordinary model. Releasing a new one could save OpenAIs bad PR right now.
A good decision amidst constant terrible ones.
Probably not financially incentivised, my guess is social climbing. He probably got reached out to by Zuck and now he’s making these posts.
Muse is nowhere near Opus 5.5.
I tested the speed of Astra and Fable 5.1 on the API and plan.
- Astra is over 50% slower compared to the API. (81 tps vs 36 tps)
- Fable speeds are identical.
Another example of how sneaky OpenAI is.