Q

[Q @Quantowski]

@thsottiaux You do realize that once you start pushing people towards API they will be more keen to experiment and maybe even switch to different providers? Are you sure you know what you are doing?

[Q @Quantowski]

@manxlab 200 raisins

[Q @Quantowski]

@JoelAbenhaim Well, how do you feel being outsmarted by beginners?

[Q @Quantowski]

@LDT0545 Usage, astra is smart enough

[Q @Quantowski]

@OpenAI Anthropic just sent me 50% discount for me to switch back to them. With 50% discount and Opus 5.5 on anthropic side and poor gpt 6 model performance it will be a tough call to be them

[Q @Quantowski]

@ah20im That's funny how they are dumber in majority of benchmarks, especially Sol

[Q @Quantowski]

@thsottiaux I hope this is a code for model that surpasses Opus 5.5

[Q @Quantowski]

@thsottiaux So is this second Codex using so much compute that it nerf usage limits and model capabilities for everyone else?

[Q @Quantowski]

@thsottiaux I hope it's Luna 6.1 and Sol 6.1 as recently released models perform poorly

[Q @Quantowski]

@GargeyaS @MiniMax_AI If it's open source you can add whatever you want

[Q @Quantowski]

@thsottiaux yup, it stopped being year of reasonable usage limits on 5x subscriptions

[Q @Quantowski]

@thdxr Or opencode GoMo (go more)

[Q @Quantowski]

@enzoatari @thdxr This is low key fire

[Q @Quantowski]

@thdxr Opencode Glide

[Q @Quantowski]

@MiniMax_AI guys, it's time to update your llm model

[Q @Quantowski]

@andrzejdragan haha, kochany Panie Profesorze, szkoda że nie wspominał Pan o tym że los matematyków będzie taki sam jak los malarzy, bo zostaną tak samo okradzeni

[Q @Quantowski]

@thsottiaux Tibo, you said the reset would be "tomorrow" so tomorrow there will be another one?

[Q @Quantowski]

@ESPNUK I want a spin off for central Europe - Cole Slav

[Q @Quantowski]

@thsottiaux I never used sub2api. What I noticed though is that the usage slowed down after 10PM central european time. So it looks like burn rate for peak hours was tripled vs previous time

[Q @Quantowski]

@thsottiaux this is so passive-agressive: give lower limits to people and at the same day present them option where they can let someone else (who maybe still have limits)finish their code