DeepSeek-V4.1-Flash-1017 just dropped - this pricing could bring the servers to their knees What do you think about the benchmarks DeepSeek has posted? Do you think it will actually live up to the numbers, or will real-world performance tell a different story? I'm really looking forward to testing it, especially at these prices. Curious to see how it performs in actual coding tasks.

New "Search Google" toggle in the Gemini tools menu I was using Gemini today and noticed a new feature when I clicked the plus icon, there is a new "Search Google" option right in the main menu. It basically lets you manually choose if Gemini should search the web for your specific prompt or not. I tried it out, and I don't know if it's that much useful. Did you guys get this update yet?

Claude opus 5.5 vs argon 4

Gemini 4 Argon usage on Pro plan compared to Opus 5.5 and GPT-6 Astra I just completed my first round of testing with Gemini 4 Argon in Antigravity. The amount of usage one gets Gemini 4 Argon on the Gemini Al PRO plan is very limited. Here is how it compares to Claude and ChatGPT plans on similar level. Preliminary number show you get about 17x the usage with Opus 5.5 and about 7x the usage with GPT-6 Astra if you have Claude Pro plan or ChatGPT Plus plan. This was tested with coding tasks. I really hope Google increases this soon. It seems to be a great model

My thoughts about argon 4 in antigravity (pro user) Google added acces to argon 4 for paid pro users and after some testing, 50% of usage dissapeared in about 12 mins of gemini thinking, and tbh i was waiting whole time for argon 4, and im simply disappointed. All those talking about fronttier and i feel like im taking with special needs ai not even joking.