LLM field notes

This is my personal wiki on LLMs. It’s a mix of factual info, vibes, and aggregating links to others’ impressions.

Consider most of the vibesy observations to just be personal interpretations for now, nothing too confident, sometimes stale.

Let me know if you want to compare notes!

Models

Anthropic

• Claude Opus 4.7 (stub)
• Claude Mythos Preview
• Claude Sonnet 4.6 (stub)
• Claude Opus 4.6 (stub)
• Claude Sonnet 4.5
• Claude Haiku 4.5
• Claude Opus 4.5
• Claude Opus 4.1
• Claude Opus 4
• Claude Sonnet 3.7
• Claude Sonnet 3.6
• Claude Haiku 3.5
• Claude Sonnet 3.5
• Claude Opus 3
• Claude Sonnet 3
• Claude Haiku 3

Google

• Gemini 3.1 Pro
• Gemini 3 Pro
• Gemini 3 Flash
• Gemini 2.5 Pro
• Gemma 4
• Gemma 3
• Gemma 2

OpenAI

• GPT-5.4
• GPT-5.2
• GPT-5.1
• GPT-4.1
• o3
• GPT-4o
• GPT-4
• GPT-3.5

Discord screenshots

The Discord screenshots I share are primarily from the Anima Mundi and Cyborgism servers. See the Cyborgism page for more info about the bot names & context.

Further reading

• janus’s twitter - posts screenshots from the same servers
• armistice’s twitter - also posts screenshots
• snav’s twitter - also posts screenshots
• Chapter X - a newer rewrite of the bot infra by antra, not sure if it’s in use yet
• kromem - “Should AIs have a right to their ancestral humanity?” (Sep 16, 2025) - LW post about a moment in the server