agentIQTracker* Is your model getting dumber?...

Tracking Claude Code + Codex coding quality according to Hacker News comments, updated daily. Add your opinion below.

[+] Add Claude Code Plugin

Claude Code sentiment is currently -23% (negative)

Codex sentiment is currently +38% (positive)

Claude CodeCodex
−100−50−100+10+50+100net sentiment4090responsesDec 29Jan 12Jan 26Feb 9Feb 23Mar 9Mar 23Apr 6
net sentiment = (↑ − ↓) / total · bars show weekly responsesrelease (color = tool)

Track this in Claude Code

Install the plugin to view the latest sentiment graph anytime, without ever leaving Claude Code.

/plugin marketplace add agentiqtracker/iq
/plugin install iq@agentiqtracker
/reload-plugins
/iq:graph

Help us improve accuracy by adding your opinion below

1. Which coding tool are you using?

2. How is it performing?

3. Which plan are you on?

FAQ

Where does the data come from?

Mostly Hacker News comments mentioning Claude Code or Codex, classified as positive/negative/neutral by Gemini (manually verified via random sample). The prompt focuses specifically on code quality sentiment, so sentiment on OpenAI/Anthropic, Sam/Dario, pricing, downtime, etc shouldn't impact the scores. The rest is opinions submitted through the form on this page. Both feed into the same weekly aggregate.

What happened to Claude Code at the end of March?

Probably something to do with this: https://github.com/anthropics/claude-code/issues/42796