# Pangram Hacker Approved page: https://hackerapproved.com/p/pangram/ 123 people recommended Pangram in Hacker News comments, 36 had mixed feelings and 161 criticized it, November 2025 to October 2026. It was recommended in 12 of 12 months, in 117 threads. It does not meet the Hacker Approved badge rule. **Verdict:** Criticized more than recommended: critics report misflagged human writing and warn against high-stakes use; those who recommend it see it as the most accurate AI text detector, with very few false positives per third-party studies. Summary written by Claude from the comments below. Claude agents reviewed it against the quotes on 2 Nov 2026. ## What people warn about - Flags genuine human writing (5 comments in the sample): "I submitted a few of my papers from grad school (long before LLMs or so-called "AI" existed) and they were flagged as being ~70% AI." ([source](https://news.ycombinator.com/item?id=48739239)) "If a tool fails the basic test of teacher getting flagged on some text that they wrote personally, that tool should never be used." ([source](https://news.ycombinator.com/item?id=49023191)) "I wouldn't trust Pangram too much. I've seen it give "100% LLM-written, high confidence" to multiple 100% human written articles" ([source](https://news.ycombinator.com/item?id=48612606)) - Unsafe for disciplining students (5 comments in the sample): "I think you can't trust Pangram in a high stakes situation, but it is absolutely better than random noise at detecting AI-generated text." ([source](https://news.ycombinator.com/item?id=49735350)) "It's good enough to be useful, but I absolutely would not be willing to fail a student on the basis of a Pangram positive." ([source](https://news.ycombinator.com/item?id=49332686)) "having 4 of them go through the "computer says you cheated, and as you can see…" ([source](https://news.ycombinator.com/item?id=49736767)) - Marketing overstates accuracy (4 comments in the sample): "if Pangram is marketing itself as "near-zero false-positives" now, that entails that the are aggressively biasing it towards being conservative" ([source](https://news.ycombinator.com/item?id=49151919)) "the accuracy figures (99.82% for Opus 5) are the true positive rate rather than a combined metric that factors in the false positive rate" ([source](https://news.ycombinator.com/item?id=49736336)) "Of course, it's unclear if the article is using the wrong term, or the wrong definition…" ([source](https://news.ycombinator.com/item?id=49021993)) - Easily fooled, misses AI text (2 comments in the sample): "Pangram is very conservative, and it's not difficult to manually get an LLM generated passage of text to turn human-written." ([source](https://news.ycombinator.com/item?id=49411137)) "Your detector did not work well on an AI-collab writing fiction project I did a while ago…" ([source](https://news.ycombinator.com/item?id=47046883)) ## Why people recommend it - Very low false positive rate (9 comments in the sample): "Their underlying classifier model is pretty accurate in my experience, though, at least in the sense that it has very few natural false positives." ([source](https://news.ycombinator.com/item?id=49213351)) "Pangram links several third party evaluations [1] that all seem to agree with 90+% detection, and very close to 0% false positives." ([source](https://news.ycombinator.com/item?id=49436120)) "pangram does have an extremely low false positive rate, which I think does make it useful for detecting cheating students" ([source](https://news.ycombinator.com/item?id=49591155)) - Confirms suspicions of AI writing (3 comments in the sample): "I'm running anything that has the "usual" AI tells through Pangram. Nine times out of ten, the article is 100% AI generated." ([source](https://news.ycombinator.com/item?id=47781868)) "I only went to Pangram because I read through the essay and could not help but notice the Claude-like sentence structure" ([source](https://news.ycombinator.com/item?id=48482768)) "if you're already thinking "this is probably AI" and pangram thinks the same, that's fairly damning" ([source](https://news.ycombinator.com/item?id=48477137)) - Backed by third-party evaluations (3 comments in the sample): "they also publish great tech reports. their founder is so confident in their model that he's regularly on social media offering bounties for false positives" ([source](https://news.ycombinator.com/item?id=49736161)) "There've been third-party evaluations of Pangram, e.g., https://bfi.uchicago.edu/wp-content/uploads/2025/09/BFI_WP_2" ([source](https://news.ycombinator.com/item?id=48130136)) "Pangram is, to my knowledge, the only AI checker today that actually passes accuracy assessments." ([source](https://news.ycombinator.com/item?id=50009249)) - Most accurate detector available (2 comments in the sample): "Pangram is fundamentally different technology, it's a large deep learning based model that is trained on hundreds of millions of human and AI examples." ([source](https://news.ycombinator.com/item?id=46089717)) "Pangram is basically the only AI detector that's actually put in the work to build an accurate classifier" ([source](https://news.ycombinator.com/item?id=49660619)) ## What people recommend instead - [GPTZero](https://hackerapproved.com/p/gptzero/): 6 recommend, 8 mixed, 27 criticize ## Where to get it - [Visit the official site](https://www.pangram.com/) ## The badge A product earns the Hacker Approved badge when at least 10 people recommend it, recommenders outnumber critics at least 2 to 1 (mixed feelings count toward neither side), it was recommended in at least 4 of the 12 months and in at least 5 different threads, and no single thread supplies more than half of the people who recommend it. Health and money products are never badged. ## Cite as Hacker Approved, "Pangram", https://hackerapproved.com/p/pangram/ (123 recommend, 36 mixed, 161 criticize; Hacker News comments, November 2025 to October 2026). Counts updated 2 Nov 2026, with data through October 31, 2026.