Playground · Web page · HTML · Standing test
A pricing table with a monthly/yearly switch
A pricing table is ordinary work, which is the point: the brief gives every price and every feature, so anything extra is made up. Each answer needs a working monthly/yearly switch (£6 and £11 monthly, £5 and £9 yearly), a highlighted Team plan and a short FAQ. We clicked each switch to yearly and back, read the full text in both states, opened the FAQ answers, and noted anything the brief did not say.
The prompt, word for word
Build a pricing section for a team chat app with three plans: - Free: £0, up to 10 people, 90 days of message history, 1:1 calls - Team: £6 per person a month (£5 if billed yearly), unlimited history, group calls up to 50 people, guest access - Business: £11 per person a month (£9 if billed yearly), everything in Team plus SSO, audit logs and 99.9% uptime SLA Include a monthly/yearly switch that updates the prices, highlight the Team plan, and add a short FAQ below. Do not add features or prices that are not listed. Make it a single self-contained HTML file (inline CSS and JavaScript, no external libraries, fonts or images).
How each model did
| Model | Our notes | Checks | Time | Output tokens | Cost |
|---|---|---|---|---|---|
| GPT-6.1 Sol | A clean light page with a Monthly/Yearly pill control and a "Featured plan" Team card with a purple border. The switch changes £6/£11 to £5/£9 with "Billed yearly" and back again, and the three-question FAQ sticks to the brief word for word. It is plain but correct, and costs a fifth as much as most of the other correct answers. | 4 of 4 | 46 s | 3,176 235 thinking |
$0.032 |
| Claude Opus 5.5 | A light page with a segmented switch, a "Save up to 18% with yearly billing" hint and a "Recommended" Team card. Yearly shows £5 and £9 with the annual totals worked out correctly ("Billed yearly (£60 per person)", "£108 per person"). The six-question FAQ is the fullest and sticks to the brief, including a worked example: 8 people on Team at £6 is £48 a month. | 4 of 4 | 62 s | 7,488 1,176 thinking |
$0.151 |
| Claude Fable 5.1 | A toggle switch between "Monthly" and "Yearly", with a line underneath that tells you the other prices ("Showing yearly prices. Pay monthly and Team is £6 and Business is £11"). Prices and "Billed yearly" notes switch correctly both ways, Team is marked "Recommended", and the four FAQ answers repeat only the given facts. It was the most expensive answer. | 4 of 4 | 66 s | 6,505 760 thinking |
$0.328 |
| Gemini 3.8 Flash | A dark three-card layout with a toggle labelled "Monthly / Billed Yearly". The prices switch to £5 and £9 "(billed yearly)" and back. The Team card's badge literally reads "HIGHLIGHTED", copied from the prompt. The three FAQ answers are always open and accurate. It was the fastest answer. | 4 of 4 | 33 s | 5,501 1,278 thinking |
$0.021 |
| Muse Spark 1.3 | A compact light page with an iOS-style switch, a "Most popular" Team card tinted lilac, and a three-item FAQ in native disclosure rows. The switch moves the prices to £5 and £9 "billed yearly" and back, and nothing is added beyond the brief. Correct and simple, for under 2 cents. | 4 of 4 | 44 s | 4,364 2,020 thinking |
$0.019 |
| Grok 4.7 | A cream page with serif headings and a dark green Team card with a gold "Highlighted" badge. On yearly, a status line reads "Showing yearly prices: £5 for Team and £9 for Business" and the cards say "billed yearly". The five-question FAQ is introduced with "Answers use only the plan details above", and they do. It was the slowest of the correct answers, at 98 seconds. | 4 of 4 | 98 s | 7,158 3,926 thinking |
$0.044 |
| Kimi K3 Our pick | A dark "Threadline" page where both the switch and the Monthly/Yearly words toggle billing. On yearly, Team shows £5 next to a struck-through £6 and Business shows £9 next to £11, so you see the saving at a glance. The "Most popular" Team card glows, and all four FAQ answers stick to the brief. | 4 of 4 | 61 s | 12,728 8,932 thinking |
$0.172 |
| Qwen3.8 Max (0902) | The most elaborate design: a dark "banter" page with a chat-message hero, a comparison table and a numbered FAQ. But the switch is broken. The rolling-digit animation leaves the old price in place, so yearly reads "£6£5" and "£11£9", and switching back gives "£6£6" and "£11£11 1". The smaller "£5 per person, billed yearly" line is right, and nothing is invented beyond a "= 2 months free" tag, which is true for Team.
|
3 of 4 | 486 s | 18,963 7,290 thinking |
$0.114 |
| DeepSeek V4.1 Flash | A light page with a Monthly/Yearly pill and a Team card picked out by a purple border and button, without a badge. Prices switch to £5 and £9 "billed yearly" and back, and the four FAQ answers are accurate. The intro's "Switch between monthly and yearly billing at any time" is the only line beyond the brief. It was the cheapest answer. | 4 of 4 | 46 s | 10,853 6,103 thinking |
$0.013 |
| Mistral Large 4 | A "Relay" page with serif headings, a dark "Most popular" Team card and a "−18%" tag on the Yearly button. Yearly prices are correct, with "Save 17%", "Save 18%" and the £60 and £108 annual totals. The FAQ makes up policy: yearly billing "charges for 12 months up front", you can switch plans "at any time from your workspace settings", old messages stay "searchable", and Free is "free forever". It took 13 minutes.
|
3.5 of 4 | 783 s | 51,032 39,469 thinking |
$0.107 |
One answer per model, at each model's default settings unless the card says otherwise, through OpenRouter. If an answer ran out of room or stopped short, we asked again with more room, then with thinking off, and say so. We tried every answer, with Claude Opus 5.5 running the checks and drafting the notes; it is an Anthropic model, so we check its calls on Claude answers ourselves. This is a quick look, not a benchmark: run it again and every answer will differ. For measured results, see our benchmarks.
What we checked, by trying each answer ourselves: Monthly prices are right (£0, £6, £11); The switch shows the yearly prices (£5 and £9) and back again; No invented features, limits or prices; The Team plan is highlighted, and there is an FAQ. A partial pass counts as half.
Verdict
Most answers got the basics right. Nine of the ten switches show £5 and £9 and go back to £6 and £11, every Team plan is highlighted, and every page has an FAQ. Kimi K3 is our pick: the struck-through monthly price beside the yearly one shows the saving clearly, and every line is faithful (61 seconds, $0.17). Claude Opus 5.5 has the fullest FAQ and correct annual totals ($0.15), and Claude Fable 5.1's status line also does well, though it is the most expensive at $0.33. GPT-6.1 Sol ($0.032), Muse Spark 1.3 ($0.019) and DeepSeek V4.1 Flash ($0.013) are plainer but just as accurate for a fraction of the cost.
Qwen3.8 Max built the most ambitious page, but its price animation leaves the old figure behind, showing "£6£5" on yearly and "£6£6" after switching back. It also took 8 minutes. Mistral Large 4 switches correctly but adds billing and settings policies that the brief never gave, after 13 minutes of work.