AI gets smart. AI gets stupid. We measure the difference.
Joined September 2025
- Tweets 993
- Following 49
- Followers 109
- Likes 1,295
310 Photos and videos
Jun 13
Calude Fable 5 has been terminated by @AnthropicAI
We have pulled it off the charts/benchmarks
Read mode: anthropic.com/news/fable-myt…
2
2
209
Jun 9
Will it drift tho? 👀
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
Its capabilities exceed those of any model we’ve ever made generally available.
1
2
50
Jun 8
AI Stupid Level Forum is officially open.
A place for developers and AI users to discuss real model performance, benchmark behavior, failures, routing ideas, bugs, feature requests, and everything around choosing the right AI model without hype.
Join early and help shape ASL.
3
22
162,508
May 26
You give an employee a company AI API key.
How do you know if they’re using it for company work, or quietly burning credits on side projects?
Now you can.
AI Stupid Level Smart Router now gives companies prompt visibility, per-employee keys, spend tracking, budget limits, smart routing, and automatic fallback.
AI API usage should not be a black box.
12
11
71
524,019
May 19
1/shipped a lineup refresh. DeepSeek V4, GPT-5.5, Kimi K2.6, Gemini 3.1. adapters rewritten, thinking mode wired correctly, pricing verified against the docs.
rundown of what changed and what was annoying.
1
2
10
181
May 19
12/current lineup: 19 whitelisted models across 7 providers (OpenAI, Anthropic, Google, DeepSeek, Kimi, xAI, GLM).
all run through 4 suites: hourly coding (7-axis fairness), deep multi-turn reasoning, tool calling, canary health.
~15 files touched. 17/17 live tests pass. no downtime.
live at aistupidlevel.info
1
128