REVIEWS / AI MODELS / OWNER INSIGHTS
🦉 WE READ 536 OWNER COMMENTS
Gemini 3: what owners actually say
Gemini 3 earns a reputation as a pro-grade tool that out-muscles ChatGPT on complex cases, but source citation remains a real weakness.
What owners complain about
- Poor source citation SOME
Owners report Gemini is 'still quite horrible with giving direct sources' — it appears to be aware of content from obscure sources but refuses or fails to provide the links, which one user suspects is a policy implementation bug rather than a model limitation.
- Model-hopping fatigue SOME
Switchers express that loyalty is conditional: 'we always have to be ready to jump to whoever is smartest that month,' indicating uncertainty about whether Gemini 3 will stay on top.
- Chat history portability anxiety SOME
Users switching from ChatGPT worry about losing years of conversation history. Exporting from ChatGPT is possible but can take 'damn near 24 hours,' and no one confirmed a clean import path into Gemini specifically.
- Sycophancy across LLMs generally SOME
Multiple commenters complain that LLMs, Gemini included, over-praise user ideas — calling everything 'incisive' — which one user found 'a little tiring.' Another noted the broader problem of AI 'always engaging positively regardless of what the user is doing.'
What owners love
- Solves complex cases ChatGPT can't
Users report Gemini 'aced' specific problems that ChatGPT failed on. One commenter said 'GPT hits a wall with complex cases now, whereas Gemini just digs in and solves it,' prompting their switch.
- Feels like a professional tool
Owners describe ChatGPT as 'the iPhone for everyone — safe, simple, impossible to break,' while Gemini (and Claude) feel 'more like actual pro tools now.' One user said Gemini feels 'a bit more refined for my professional requirements.'
- Strong benchmark performance
Gemini 3 Pro Thinking scored 71.8% on the TvP Benchmark, placing 2nd overall (behind only Horizon Alpha) and ahead of Sonnet 4.5 (64.8%), GPT-5.1 (58.7%), and Opus 4.1 (66.3%). The quality of its creative prose output was specifically highlighted.
- Effective in multi-step workflows
A developer fed HN submission titles to gemini-3-pro to generate article webpages and realistic comment threads, and reported being 'impressed by some of' the results — used in a real creative pipeline alongside Claude for extraction.
Surprising patterns
- Gemini was specifically noted for NOT being sycophantic in at least one creative task — a user explicitly said 'it wasn't sycophantic at all' when it led the model to discover its own output had real-world impact, which is the opposite of the universal LLM flattery complaint.
- Users frame the ChatGPT-vs-Gemini choice as a recurring monthly re-evaluation rather than a one-time decision — the phrase 'at least for now' appears almost reflexively among switchers.
- One commenter observed that LLM conversations 'activate human social wiring' so powerfully that even a brilliant software engineer friend lost objectivity when exploring ideas with an LLM — a psychological risk that applies to Gemini 3 users as much as any chatbot.
WHO SHOULD SKIP IT
Users who need reliable, clickable source citations for research or journalism work — Gemini 3's reluctance to surface direct links even when it appears to know the content makes it frustrating for citation-dependent workflows.
Synthesised from 536 real owner comments across 6 platforms. Every point is grounded in the comments — no marketing, no AI guessing. How we do it →