anthropic vs openai models - simplify for us please

every other day i feel like im reading some biased reddit article or misinformation from either openai or anthropic fan boys claiming how much better their respective models are compared to the competition... its exhausting honestly. ive been using openai based models for 2 months now, mainly for chatgpt, hermes managing my server and t3 codes coding agent (for fully vibe coded apps). i use mostly gpt 6 luna on high and by the end of the week ive basically still got 95% usage left so im not an intensive user. accroding to livebench.ai and artificialanalysis.ai/models i think openai models are way more cost effective and you seem to get a smarter model for less money is this true? has anyone got general tips for filtering out the fan boy reviews and getting a better idea for whats actually better. the main thing im looking for is the best value for money from my respective model, whilst my usage is not intensive yet, it will pick up a bit pretty soon any advice appreciated, thanks