So OpenAI finally dropped GPT-5 after all the hype and that weird Death Star tweet from Sam Altman. I’ll be real with you guys - I had massive expectations for this release. Like, I was expecting some groundbreaking stuff that would blow my mind.
But after using it for a few days now… it’s just not hitting the way I thought it would.
Don’t get me wrong, it’s definitely an improvement over GPT-4, but the jump feels more like going from a Honda Civic to a Honda Accord rather than upgrading to a Tesla, you know what I mean?
Here’s my take after testing it extensively:
For coding specifically, I still think Anthropic’s Claude (both Opus and the newer Sonnet 4) absolutely destroys GPT-5. The difference is honestly night and day. Claude just gets what I’m trying to build, understands the context better, and gives me solutions that actually work without needing 5 iterations.
For general creative writing tasks, there’s maybe like a 15-20% difference between them, but even there Claude feels more… human? Like it actually understands what I’m trying to convey instead of just pattern matching.
What really gets me is the context understanding. Claude seems to read between the lines and figure out what I actually want to achieve, not just what I literally typed. GPT-5 still feels a bit rigid in comparison.
Maybe I’m being too harsh since it just launched, but after waiting this long I expected something more revolutionary rather than evolutionary.
What’s your experience been? Are you seeing the same things or am I missing something here? Has anyone done proper side-by-side comparisons yet?
Really curious to hear what others think, especially those doing serious coding work or creative projects.
I use gemini 2.5 for coding and it works excellent. The biggest plus with gemini is that the context window is super large. But yes, Claude works great too.
I reckon it is the plateauing of the progress chart, every AI model sort of performs like every other model for 80% of the tasks. Now enthusiasts will definitely find differences, but it feels like the AI progress is following the smartphone progress curve.
One thing is plenty clear though, largely the innovation is coming from China and the western counterparts are playing catch up. Genie 3 was an exception, it was legitimiately mind blowing.
What would you guys use for deep research on financials, reading between the lines..?
currently using Grok 4 and Perplexity..seems to be fine but I want something that is way superior and capable..
AI is such a broad topic but it’s same as dot.com, AI is massively over rated in my opinion, and it isn’t naturally evolving, in fact, it has a saturation point, we just don’t have any more data to feed to improve it, telling everyone would lose their jobs , from my point of view, is such a mehh statement, .. developer does more than things that only includes coding..but nevertheless , GPT really made my life easier.. and GPT 5 seems good but it’s not like so extraordinary or life changing