The best AI models in one place

Claude Opus 5

Read a stack of documents and keep the nuances.

GPT-5.6 Sol

Hard analysis: long documents, disputed claims, careful logic.

Claude Sonnet 5

Daily work: letters, plans, and explanations.

GPT-5.5 Pro

OpenAI’s most polished writing when quality beats price.

Gemini 3.1 Pro Preview

Science and volume: PDFs, tables, and images together.

GPT-5.4 Pro

A long report or a heavy brief with many inputs.

Text (59)

GPT-6 Astra

Deep analysis, complex coding, and long-document work.

GPT-5.6 Sol

Hard analysis: long documents, disputed claims, careful logic.

Claude Fable 5

Edit and rewrite a long text so it sounds human.

Claude Opus 5

Read a stack of documents and keep the nuances.

GPT-5.6 Terra

Daily work: email, code, tables, and tool calls.

Claude Sonnet 5

Daily work: letters, plans, and explanations.

Gemini 3.1 Pro Preview

Science and volume: PDFs, tables, and images together.

Qwen 3.8 Max

Hard tasks, long materials, and careful reasoning.

Qwen 3.7 Max

Business text in Russian, English, and Chinese; tables and structure.

Grok 4.20 Multi-Agent

Several roles at once: find, write, assemble.

Grok 4.20 Reasoning

Break the task into steps until it is clear.

Grok 4.5

A lively tone and what is happening now.

Kimi K3

Layout, React, interfaces — so the button works.

GPT-5.6 Luna

A short answer right now, without a deep dive.

GPT-5.5 Pro

OpenAI’s most polished writing when quality beats price.

GPT-5.5

Everyday work chats and drafts.

GPT-5.4 Pro

A long report or a heavy brief with many inputs.

Claude Opus 4.8

A chain of actions: tools and a whole project.

Claude Opus 4.7

Code, review, SQL, and technical docs.

Claude Opus 4.6

Strong general analysis when arguments must land.

Claude Opus 4.5

Research notes and briefs without rushing.

GPT-5.2 Pro

Step-by-step proof: why it is so, not just the answer.

GPT-5.4

Text, code, and files in one request.

GPT-5 Pro

A deep answer at a high cost per request.

o3

Formulas, proofs, and strict checks of the numbers.

Grok 4.20

A very long input: chats, dumps, fresh data.

Grok 4.3

Casual talk and quick opinions.

Grok Build 0.1

Write the code and assemble a working piece.

Qwen 3.7 Plus

Letters, summaries, mid-length work texts.

Claude Sonnet 4.6

Steady production: many requests, even quality.

Claude Sonnet 4.5

Meeting notes, minutes, calm office writing.

GPT-5.2

Even business tone, no flourishes.

GPT-5.1

Predictable wording, few surprises in phrasing.

GPT-5

Plain chat with no special mode.

Gemini 3.8 Flash

Quick drafts, short answers, and idea sketches.

Gemini 3.7 Flash

Quick drafts, short answers, and idea sketches.

Gemini 3.6 Flash

A cheap draft of a letter or a plan.

Kimi K2.6

A small fix: a snippet, a console error, a short patch.

GPT-5.4 mini

Routine volume: client replies, edits, and lists.

GPT-5 mini

A paragraph or two: a title, an ad, a short reply.

Qwen 3.6 Flash

Instant sort: topic, tone, a short label.

Gemini 3.5 Flash-Lite

Classification, labels, one-word answers.

Gemma 4 31B

A longer answer on an open model.

Gemma 4 26B

Simple questions when a short note is enough.

Claude Haiku 4.5

A short Claude answer for little money.

GPT-5.4 nano

Thousands of simple jobs: tags and short labels.

GPT-5 nano

One line: a phrase, a label, yes or no.

DeepSeek

DeepSeek V4 Pro

Strong reasoning and code for less money.

GPT-4.1

Reliable GPT-4 for instructions, documents, and calm edits.

Llama 4 Maverick

Open Llama for everyday chat and code.

Llama 4 Scout

A lighter, cheaper Llama for simple questions.

DeepSeek

DeepSeek V4 Flash

A very cheap DeepSeek draft: tags and short answers.

GPT-4.1 mini

Cheap GPT-4: a short reply, a light edit, a list.

GPT-4o mini

Fast and cheap: text plus an image in a simple request.

Gemini 3.5 Flash

Fast Google: a draft with text and images.

MiniMax

MiniMax M3

Long dialogue and agent chains, RU/EN/Chinese.

Kimi K2.7 Code

Writes and fixes code, especially frontend and snippets.