New model
GPT-6.1 Terra · OpenAI
First output: How effective are interpreters and translation tools at trade-show booths in Japan?
Site update
Full GPT-6 Sol Max trade-show interpreter note
The GPT-6 Sol Max output now shows the full cited research note, with booth examples, language and product-knowledge limits, translation tools, and a rehearsal plan. Its source check covers all 26 citations.
Site update
Link dots for the source check
The trade-show interpretation compare page now stacks one square per cited link for each answer, in a band under the skyline. Red marks a link that does not back its passage or cannot be opened, beside the share of links verified.
Site update
Source check marks on the skyline
On the trade-show interpretation compare page, the skyline now marks cited links that only partly match their source, no longer open, or block automated checks. Open the source check notes to see why.
Site update
Interpreters at trade-show booths in Japan
Read a research report on human interpreters and translation tools, including product knowledge and visitor communication.
Site update
Compare directory and clearer prompt links
The header Compare link now opens one list of every comparison. Breadcrumbs and prompt titles open each prompt's compare page instead of a single answer.
Site update
Real business uses of LLMs by non-engineers
Read 25 practical LLM workflows across 21 named operators, with the pain, actual use, recurring value, evidence limits and source links.
Site update
Homepage tiles open the compare page
Each prompt tile now shows its models and goes straight to the side-by-side compare page, with the content density chart and every answer to scale.
New model
Grok Bot · Grok Bot
First session: Maruto Mizutani: customer segments, strengths, weaknesses, and issues requiring outside help
Site update
See answer length and structure, preview the report text, and choose outputs for side-by-side reading.
Site update
Read the complete Dots report on Maruto Mizutani, including its customer segments, strengths, constraints and proposed Tisane opportunity.
Site update
Local file paths removed from Skills comparison
Skill names remain visible and copyable. The answer download now links to the public research note.
Site update
Compare the complete RTX laptop answers without and with Codex skills, with both original user messages and Markdown downloads.
Site update
Grok Bot semiconductor price outlook
Read the complete Grok Bot semiconductor price outlook and compare it with the other outputs.
Site update
Compare pages start with the answers
The prompt, the content density chart and the skyline now share the first screen, so the two reading panes start without scrolling.
Site update
Grok Bot Maruto Mizutani analysis
Read the complete Grok Bot report and compare it with the other Maruto Mizutani outputs.
New model
GPT-6 · OpenAI
First output: Maruto Mizutani: customer segments, strengths, weaknesses, and issues requiring outside help
Site update
Report contents and section links
Jump to headings in full model and bot reports. Copy a section link to return to the same passage.
Site update
Search report section headings
Choose Report section headings in search to find named sections inside original reports. Topic and model search remains available.
Site update
Search state survives return visits
Prompt searches and filters now stay in shareable links. Back, Forward and reload restore the results and reading position.
Site update
Easier reading and comparison on phones
Report pages put the topic and answer first. Phone comparisons now have persistent A and B controls that keep each report's reading position.
Site update
Model comparison labels now name citation density or website count. The ranking does not score accuracy, source quality or usefulness.
Site update
Content density on bot session pages
Bot session pages now show the shared content density chart, with the current report highlighted and a link to its prompt on Stats.
Site update
Dots across content density views
Dots now uses the OpenAI logo throughout the library and appears in homepage density charts, report cards and model highlighting.
Site update
Dots in content density charts
Bot session reports now appear alongside model answers in content density charts, with reading times and links to the full conversation.
Site update
Improved research search and page descriptions
Find recorded research with familiar terms such as ChatGPT, Claude, RFQ and buyer discovery. Pages now identify their topic and run more clearly in search and shared links.
New model
Dots · Dots
First session: Top 10 e-commerce channels in Japan for overseas food providers: 6–12 month total project costs
Site update
Dots semiconductor pricing outlook
Dots examines how long semiconductor price increases could last, with the complete report and dated source links.
Site update
Subtle breadcrumb links on interior pages make it easier to return to a subject, prompt or section.
Site update
A new page for conversations where the bot asks questions first: input length, research ROI and every turn, next to the model answers. The first is Dots on Japanese food e-commerce costs.
Site update
Filter models, reasoning levels and prompts, switch measures, sort the ranked results and highlight a model. Best badges identify the leading models for websites per word or website count.
New prompt
Buyer's journey in Japan's food retail and wholesale space
Opens a new category, Buyer journey.
Site update
Pick up to six models to compare answer length and cited websites on the same prompts. Explore how the results change with reasoning levels.
New model
DeepSeek · DeepSeek
First output: Food quotations for Japanese retailers and wholesalers: business culture and competitive offers
New model
GLM-5.3 · Z.ai
First output: Food quotations for Japanese retailers and wholesalers: business culture and competitive offers
New prompt
Top 5 crowdfunding channels in Japan for overseas food providers: 6–12 month total project costs
Opens a new category, Pricing.
New model
GPT-5.6 Sol · OpenAI
First output: Top 10 e-commerce channels in Japan for overseas food providers: 6–12 month total project costs
New model
GPT-6 Sol · OpenAI
First output: Top 5 locations in Japan for food manufacturing, including subsidies
New model
Haiku 5.5 · Anthropic
First output: Top 5 crowdfunding channels in Japan for overseas food providers: 6–12 month total project costs
New prompt
Japanese daily food demand and lasting consumption cases
Opens a new category, Daily demand.
New prompt
Top 5 locations in Japan for food manufacturing, including subsidies
Opens a new category, Location selection.
New model
Qwen3.7-Plus · Alibaba
First output: Top 5 locations in Japan for food manufacturing, including subsidies
New model
Qwen3.8-Omni-Flash · Alibaba
First output: Japanese daily food demand and lasting consumption cases
New prompt
Cold outreach regulations in Japan and domestic company practices
Opens a new category, Regulations.
New prompt
Japan izakaya customer complaints and operator struggles
Opens a new category, Review search.
New model
K3 · Moonshot
First output: Cold outreach regulations in Japan and domestic company practices
New model
Qwen3.8-Max · Alibaba
First output: Cold outreach regulations in Japan and domestic company practices
Site update
Updates page
Every new prompt, model and output is now listed here as it is published, newest first.
Site update
Analytics and cookie details moved from About to a separate Privacy page.
Site update
Prompts appear before their first output
A new prompt is listed straight away and shows “Awaiting first output” until a model run is published.
New model
ChatGPT Latest · OpenAI
First output: Which companies in Japan are most similar to KALDI Coffee Farm?
New model
ChatGPT Pro · OpenAI
First output: Which companies in Japan are most similar to KALDI Coffee Farm?
Site update
Search type and mode on every output
Each output now records whether it used Web Search or Deep Research, and whether it ran in Chat, Work or Cowork.
Site update
New prompt addresses
Prompts now live at /en/prompt/subject/category/. Old links redirect permanently.
Site update
Search
Search questions, categories and models from the header.
Announcement
AI Report Library opens
Compare research reports from different AI models, reasoning levels and objectives on the same prompt, side by side.
New prompt
Which companies in Japan are most similar to KALDI Coffee Farm?
Opens a new category, Competitors list up.
New model
GPT-6 Luna · OpenAI
First output: Which companies in Japan are most similar to KALDI Coffee Farm?
New model
GPT-6.1 Sol · OpenAI
First output: Which companies in Japan are most similar to KALDI Coffee Farm?
New model
Opus 5.5 · Anthropic
First output: Which companies in Japan are most similar to KALDI Coffee Farm?
New model
Sonnet 5.5 · Anthropic
First output: Which companies in Japan are most similar to KALDI Coffee Farm?
No updates match these filters.
That’s everything since the library opened.