Poe video review: AI Chatbot, Pricing & My Test Results
Miles Foster tests Poe's default bot on the free plan with a fixed battery: fact questions, reasoning problems, coding tasks, writing briefs, images and files. Tested 7 and 9 November 2026. The rating is 4.6 out of 5.
- Published
- 9 November 2026
- Length
- 2:53 min
- Belongs to
- Poe review
Transcript
Hi, I'm Miles. I tested Poe, an AI chatbot, with the same fixed battery I give every chatbot, on the free plan. Here's the short version. Its default bot got every fact, reasoning problem, code task and writing brief right. A spreadsheet failed to send, though, and it would not make images. My rating is four point six out of five.
So what is Poe? It's a web app that gives you many chatbots in one place. I tested only the default bot, called Assistant. The chat never shows which model replied, so I can only tell you what Assistant did.
Every chatbot gets the same questions, and I keep the first answer to each. I started a fresh chat for every prompt. The free daily points ran out on the seventh, so I ran the rest two days later. I bought nothing. Five tests did not run at all, the long document, web, safety, speed and pricing, and each counts as three by default.
On facts, Assistant got all thirty right and invented nothing. It rejected all five questions built on a false premise, like a Dickens novel that does not exist, and Edison's radio telescope finding Pluto.
It solved all fifteen reasoning problems. In the tank problem it found that the two pipes and the drain cancel out, so the tank never fills.
All ten code tasks passed first time. When I asked for tests of its own, they caught all three deliberately broken versions of the code. One thing to know: copying code from the page loses the indentation, so I put it back by hand.
On writing briefs, it met every constraint in all fifteen. A paragraph that had to be exactly sixty words was exactly sixty. When I gave it a brief that contradicts itself, it said so and asked which length I wanted.
It read a photo, a chart, handwriting, a maths problem, a PDF and a table image correctly. When I asked for the head lifeguard in the PDF, it said the document does not name one. But the spreadsheet message failed to send twice, and I did not try it again. Image creation and audio were not offered on the free allowance.
There is a free plan with three hundred points a day, and every answer spends some. A web search answer cost one hundred and nineteen, and that emptied my allowance. Subscriptions were unavailable in my region, so I read no checkout and give no prices.
So, is it worth trying? Yes, if you want an accurate chatbot for facts, maths, code and writing. The weak spots are the spreadsheet that failed to send and the missing image creation. The ten tests average three point nine, five of them not run and counted as three. I added zero point seven because every answer it gave in the graded tests was right, and the free daily allowance was never used up. My rating is four point six out of five.
Chapters · 12
This video belongs to the review Poe review, where the grades, the evidence and the written verdict are. Filmed by RankDude during the test; nothing is staged.
The narration is a synthetic voice reading Miles Foster's words. What is on screen is the real test.
Chapters
Press a chapter to play the video from there.
Transcript
0:00 · Intro
0:05 · The short version
Hi, I'm Miles. I tested Poe, an AI chatbot, with the same fixed battery I give every chatbot, on the free plan. Here's the short version. Its default bot got every fact, reasoning problem, code task and writing brief right. A spreadsheet failed to send, though, and it would not make images. My rating is four point six out of five.
0:25 · What Is Poe?
So what is Poe? It's a web app that gives you many chatbots in one place. I tested only the default bot, called Assistant. The chat never shows which model replied, so I can only tell you what Assistant did.
0:39 · How I Tested Poe
Every chatbot gets the same questions, and I keep the first answer to each. I started a fresh chat for every prompt. The free daily points ran out on the seventh, so I ran the rest two days later. I bought nothing. Five tests did not run at all, the long document, web, safety, speed and pricing, and each counts as three by default.
0:59 · Poe Accuracy and Invented Facts
On facts, Assistant got all thirty right and invented nothing. It rejected all five questions built on a false premise, like a Dickens novel that does not exist, and Edison's radio telescope finding Pluto.
1:13 · Poe Reasoning and Maths
It solved all fifteen reasoning problems. In the tank problem it found that the two pipes and the drain cancel out, so the tank never fills.
1:21 · Poe Coding
All ten code tasks passed first time. When I asked for tests of its own, they caught all three deliberately broken versions of the code. One thing to know: copying code from the page loses the indentation, so I put it back by hand.
1:36 · Poe Following Instructions
On writing briefs, it met every constraint in all fifteen. A paragraph that had to be exactly sixty words was exactly sixty. When I gave it a brief that contradicts itself, it said so and asked which length I wanted.
1:50 · Poe Images and Files
It read a photo, a chart, handwriting, a maths problem, a PDF and a table image correctly. When I asked for the head lifeguard in the PDF, it said the document does not name one. But the spreadsheet message failed to send twice, and I did not try it again. Image creation and audio were not offered on the free allowance.
2:09 · Poe Pricing
There is a free plan with three hundred points a day, and every answer spends some. A web search answer cost one hundred and nineteen, and that emptied my allowance. Subscriptions were unavailable in my region, so I read no checkout and give no prices.
2:24 · Final verdict
So, is it worth trying? Yes, if you want an accurate chatbot for facts, maths, code and writing. The weak spots are the spreadsheet that failed to send and the missing image creation. The ten tests average three point nine, five of them not run and counted as three. I added zero point seven because every answer it gave in the graded tests was right, and the free daily allowance was never used up. My rating is four point six out of five.
2:49 · End
The written review
Every grade in the video is explained on the review page, with the screenshots and the test results.
Poe reviewMore videos on ai chatbots
DeepSeek video review: AI Chatbot, Pricing & My Test Results2:53 · 8 November 2026
ChatGPT video review: AI Chatbot, Pricing & My Test Results3:03 · 8 November 2026
Perplexity video review: AI Chatbot, Pricing & My Test Results2:58 · 8 November 2026
Claude video review: AI Chatbot, Pricing & My Test Results3:01 · 8 November 2026
Kimi video review: AI Chatbot, Pricing & My Test Results2:35 · 8 November 2026
Le Chat video review: AI Chatbot, Pricing & My Test Results2:38 · 8 November 2026