π OpenAI just dropped GPT-4.1! π Now available in three sizes: GPT-4.1, GPT-4.1 Mini, and GPT-4.1 Nano. The 4.1 and Mini have been on a test drive for weeks (codenamed Quasar and Optimus) on the arena and openrouter, and theyβre already rocking the API! What's new? π€ Everything got a glow-upβfrom coding to multimodal features! But the standout? 4.1 plays nice with instructions way better, so no more forgetting what you asked and definitely no more gaslighting. π
ββοΈ All models now support a whopping million tokens in context (at least in the API), and guess what? Longer context wonβt cost you extra! π° But how this will roll out for ChatGPT users is still a mysteryβPlus/Team subscribers are stuck at 32k tokens while free users only get 8k. Oh, and hereβs the kicker: Full-on 4.1 is 20% cheaper per token than GPT-4o! However, the Mini is pricier than the old Mini version, and Nano isnβt quite ready to take its placeβit lags behind in some benchmarks. πββοΈπ¨ Plus, cached token dis