# OpenAI GPT 1205 Hacker News commenters recommended OpenAI GPT and 987 criticized it, November 2025 to November 2026. It was recommended in 13 of 13 months; owners who said how long report a median of 0.5 years (47 owners). **Verdict:** Recommended for strong coding, debugging and research help with good reasoning and token efficiency, though critics cite hallucinations, overcomplication, verbose output and high cost. ## Why people recommend it - Strong at debugging and engineering (5 comments): "I've found GPT/Codex models consistently stronger in engineering tasks (such as debugging complex, cross-systems issues, concurrency problems, etc)." ([source](https://news.ycombinator.com/item?id=47707836)) "I do prefer the overall results with GPT 5.4, which seems to catch more bugs in reviews that Gemini misses and produce cleaner code overall." ([source](https://news.ycombinator.com/item?id=47864717)) "GPT 5.5 is a good example of a model that's generally better at solving problems than Opus 4.8, but the code it generates is worse" ([source](https://news.ycombinator.com/item?id=48828605)) - Handles long autonomous tasks (4 comments): "GPT 5.5 which quite literally worked through the night day (with the /goal feature), and actually had something valuable and useful in the end" ([source](https://news.ycombinator.com/item?id=48608235)) "it can walk away for 45 minutes, produce CoT with a length of Empire State Building and give me fairly good and well written response" ([source](https://news.ycombinator.com/item?id=46606229)) "I use gpt 5.5 high with planning every time and plan various smaller features/changes in parallel, then approve them one after another." ([source](https://news.ycombinator.com/item?id=48815047)) - Token efficient, less chatty (2 comments): "Most variants of GPT-5.5 are less chatty and token-intensive than Opus 4.8/4.7, so despite the output token price being higher, it generates fewer tokens" ([source](https://news.ycombinator.com/item?id=48315007)) "it does seem like a step change in token efficiency, though based on the earlier artificial analysis reporting it's also quite the cost lottery" ([source](https://news.ycombinator.com/item?id=48061483)) - Accurate research and expert answers (5 comments): "I used GPT-4 as a second opinion on my medical tests and doctor's advice, and it suggested an alternate diagnosis and treatment plan that turned out to be correct." ([source](https://news.ycombinator.com/item?id=45836692)) "answer all my tax questions correctly. Those thousands of dollars in time and money I had spent has been replaced by a single $20 subscription." ([source](https://news.ycombinator.com/item?id=49224992)) "end up with a report indistinguishable in accuracy and bibliographic quality from what a good human might produce in a couple of weeks" ([source](https://news.ycombinator.com/item?id=46722888)) ## What people warn about - Hallucinates tools and features (3 comments): "hallucinated the existence of two CLI utilities, and when corrected, then hallucinated the existence of non-existent, but plausible, features/options" ([source](https://news.ycombinator.com/item?id=46520569)) "It was going in loops hallucinating features of GCP, looking things up in strange ways, running terraform apply after being explicitly told not to" ([source](https://news.ycombinator.com/item?id=48510803)) "4.1 made nonsense changes" ([source](https://news.ycombinator.com/item?id=46542620)) - Verbose, too many options (4 comments): "GPT can be very creative when discussing a high-level design. Once it starts writing code snippets it will offer to take things in a bunch of different directions." ([source](https://news.ycombinator.com/item?id=48587199)) "Receiving a half-baked, extremely verbose recap of thinking on benign details with Opus 4.8 or GPT 5.5 feels like an extraordinary loss of quality of experience" ([source](https://news.ycombinator.com/item?id=48852273)) "GPT's output is awful and it gets even more awful when you try to work out a solution "together" because it shits out 10 paragraphs with 20 options" ([source](https://news.ycombinator.com/item?id=47849319)) - Expensive to use (3 comments): "Anecdotally, GPT 5.5 seems decent, but it's rumored to be a 10T parameter model, isn't noticeably better than 5.4 or 5.3, is insanely expensive to use" ([source](https://news.ycombinator.com/item?id=48015864)) "which in my opinion are unlikely to have been used for anything serious, as they're simply several times more expensive than any human effort" ([source](https://news.ycombinator.com/item?id=48256578)) "It's a cash grab. More conversational AI means more folks running out of free or lower paid tier tokens faster, leading to more upsell opportunities." ([source](https://news.ycombinator.com/item?id=45916811)) - Overcomplicates or misses big picture (4 comments): "5.5 is overcomplicating it. Where the solution is e.g. changing some oidc auth url, it goes around and verifies and check and builds this and that" ([source](https://news.ycombinator.com/item?id=48076281)) "GPT-5 models are much more fixated on solving exactly the problem given and not as good at taking a step back and thinking about the big picture" ([source](https://news.ycombinator.com/item?id=46542555)) "GPT-5.3+ are some of the smartest models out there and could be a pleasure to work with, if they weren't lazy bastards to the point of being completely infuriating." ([source](https://news.ycombinator.com/item?id=47895295)) ## Alternatives hackers recommend - [Claude Opus](https://hackerapproved.com/p/claude-opus.md): 3331 recommend - [Claude Sonnet](https://hackerapproved.com/p/claude-sonnet.md): 854 recommend - [DeepSeek V4 Flash](https://hackerapproved.com/p/deepseek-v4-flash.md): 828 recommend - [Fable](https://hackerapproved.com/p/fable.md): 823 recommend - [Claude Fable 5](https://hackerapproved.com/p/fable-5.md): 774 recommend ## Where to get it - [Visit the official site](https://openai.com/) ## Cite as Hacker Approved, "OpenAI GPT", https://hackerapproved.com/p/gpt-5-5/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).