8 ms·
Even the "low" pelican is better than most other models.
by sanex 12d ago
Even the "low" pelican is better than most other models.
- matt_heimer 12d agoIt's subjective but I think only gpt-5.6-sol XHIGH is comparable to gpt-6-astra LOW in quality. And it's 24.11 cents vs 9.55 for astra low.
- sanex 12d agoSomething I've noticed especially with fable at work is it seems to be smarter, so it takes fewer tokens and burns less of my usage than smaller models would for similar tasks.