PodcastIntelligence Snacks · Weekly conversations about AI, software and business, from the team behind Normie Mode.Listen →
Best AI for…

Best AI for documents, PDFs and notes

Reading PDFs and documents, pulling out information and taking notes.

Best overall

Claude Fable 5.1

Get it with Claude Max, $100 a month (for up to half your weekly allowance)

Top of 13 models for this task

Best for $25 a month or less

Gemini 3.8 Flash

Get it with Google AI Pro, $19.99 a month

2nd of 13 models for this task

Best free

Gemini 3.1 Pro

Get it with Gemini Free (limited use)

3rd of 13 models for this task

Keep in mind Covers pulling data out of documents; nothing yet tests summaries or meeting notes.

Based on tests of this task. The picks follow a fixed rule using the charts below and each app's plans.

Already paying for one? See what each plan gives you here
App and planPriceBest model on it for thisPosition
ChatGPT FreefreeGPT-5.6 Luna11th of 13
ChatGPT Go$8 a monthGPT-5.6 Luna11th of 13
ChatGPT Plus$20 a monthGPT-5.6 Sol4th of 13
ChatGPT Pro$100 a monthGPT-6 Astra
Called GPT-6 Pro in the app
7th of 13
Claude FreefreeClaude Sonnet 59th of 13
Claude Pro$20 a monthClaude Sonnet 5
Its main model, Claude Opus 5.5, is too new to score yet.
9th of 13
Claude Max$100 a monthClaude Fable 5.1
For up to half your weekly allowance
Its main model, Claude Opus 5.5, is too new to score yet.
1st of 13
Gemini FreefreeGemini 3.1 Pro
Limited use
3rd of 13
Google AI Plus$4.99 a monthGemini 3.1 Pro
Twice the free plan's use
3rd of 13
Google AI Pro$19.99 a monthGemini 3.8 Flash2nd of 13
Google AI Ultra$99.99 a monthGemini 3.8 Flash2nd of 13
Kimi FreefreeKimi K36th of 13
Kimi Moderato$19 a monthKimi K36th of 13

Norm's guide

What I'd tell a friend who asked.

What AI is good at

  • Summarising long documents.
  • Pulling figures and names into a table.
  • Turning meeting notes into action points.

Where it trips up

  • Skipping parts of very long documents.
  • Mixing up similar figures or names.
  • Scanned documents where the image is poor.

How to get a good result

  • Ask it to quote the part of the document it's using.
  • Spot-check a few details against the original.
  • For long files, ask about one section at a time.

How the models compare

Every model we could score for documents, PDFs and notes, combining the tests below.

Overall for documents, PDFs and notes

Position on a combined score from 2 tests. The longer the bar, the further ahead · Higher is better

On the tests we have, Claude Fable 5.1 comes out on top, followed by Gemini 3.8 Flash and Gemini 3.1 Pro.

Claude Fable 5.1 · Anthropic1st
Gemini 3.8 Flash · Google DeepMind2nd
Gemini 3.1 Pro · Google DeepMind3rd
GPT-5.6 Sol · OpenAI4th
Gemini 3.6 Flash · Google DeepMind5th
Kimi K3 · Moonshot AI6th
GPT-6 Astra · OpenAI7th
DeepSeek V4 Pro · DeepSeek8th
Claude Sonnet 5 · Anthropic9th
DeepSeek V4.1 Flash · DeepSeek10th
GPT-5.6 Luna · OpenAI11th
Gemini 3.5 Flash-Lite · Google DeepMind12th
Claude Haiku 4.5 · Anthropic13th

The evidence

Pulling information out of documents into tables

Tests this task · Reading documents and filling in a table with the right figures, including cases that need reasoning. · Higher is better

Claude Fable 5.1 leads, followed by GPT-6 Astra and Gemini 3.1 Pro.

Claude Fable 5.1 · Anthropic98%
GPT-6 Astra · OpenAI97%
Gemini 3.1 Pro · Google DeepMind97%
GPT-5.6 Sol · OpenAI96%
Gemini 3.8 Flash · Google DeepMind96%
Gemini 3.6 Flash · Google DeepMind95%
DeepSeek V4 Pro · DeepSeek94%
Claude Sonnet 5 · Anthropic92%
Kimi K3 · Moonshot AI91%
DeepSeek V4.1 Flash · DeepSeek90%
GPT-5.6 Luna · OpenAI89%
Gemini 3.5 Flash-Lite · Google DeepMind83%
Claude Haiku 4.5 · Anthropic74%
Source: DTBench authors via Epoch AI · CC BY 4.0

People's votes on long, detailed requests

Tests a related skill · People compared two anonymous answers to long requests with lots of detail to take in. · Higher is better

Claude Opus 5.5 and Claude Fable 5.1 are neck and neck at the top, followed by Gemini 3.8 Flash and Kimi K3; Claude Fable 5.1 costs about 3 times as much.

Claude Opus 5.5 · Anthropic1st
Claude Fable 5.1 · Anthropic2nd
Gemini 3.8 Flash · Google DeepMind3rd
Kimi K3 · Moonshot AI4th
Gemini 3.1 Pro · Google DeepMind5th
GPT-5.6 Sol · OpenAI6th
DeepSeek V4.1 Flash · DeepSeek7th
Gemini 3.6 Flash · Google DeepMind8th
Claude Sonnet 5 · Anthropic9th
GPT-6 Astra · OpenAI10th
DeepSeek V4 Pro · DeepSeek11th
GPT-5.6 Luna · OpenAI12th
Gemini 3.5 Flash-Lite · Google DeepMind13th
Claude Haiku 4.5 · Anthropic14th
Source: LMArena · votes as of 25 Sept 2026

If you build with it

What the makers charge developers who use the models directly. Using an app? The plans above are what you pay.

Cost per 1,000 typical requests

List price, about 1,500 words in and 500 out per request · Lower is better

GPT-6 Luna is cheapest, followed by DeepSeek V4.1 Flash and GPT-5.6 Luna.

GPT-6 Luna · OpenAI$0.55
DeepSeek V4.1 Flash · DeepSeek$0.72
GPT-5.6 Luna · OpenAI$1.24
DeepSeek V4 Pro · DeepSeek$1.48
Gemini 3.5 Flash-Lite · Google DeepMind$2.35
Gemini 3.8 Flash · Google DeepMind$4.13
Gemini 3.6 Flash · Google DeepMind$4.13
Claude Haiku 4.5 · Anthropic$5.50
Claude Sonnet 5 · Anthropic$11
GPT-6 Sol · OpenAI$11
Gemini 3.1 Pro · Google DeepMind$12
Kimi K3 · Moonshot AI$17
GPT-5.6 Sol · OpenAI$22
Claude Opus 5.5 · Anthropic$22
GPT-6 Astra · OpenAI$55
Claude Fable 5.1 · Anthropic$55
Source: models.dev · MIT
Intelligence Snacks newsletter

The big AI ideas each week, in your inbox