MaralGPT "Mythos" 9B just released, and this is why I'm proud of my project
haghiri75
11 points
12 comments
June 29, 2026
Related Discussions
Found 5 related stories in 710.1ms across 14,015 title embeddings via pgvector HNSW
- GPT-5.5: Mythos-Like Hacking, Open to All rs_rs_rs_rs_rs · 53 pts · April 23, 2026 · 52% similar
- What it feels like to work with Mythos swolpers · 223 pts · June 09, 2026 · 51% similar
- A quick look at Mythos run on Firefox: too much hype? leonidasv · 48 pts · April 24, 2026 · 47% similar
- Mythos is the best cybersecurity news in a decade littlexsparkee · 42 pts · May 06, 2026 · 47% similar
- Will It Mythos? mindingnever · 73 pts · June 23, 2026 · 46% similar
Discussion Highlights (8 comments)
vanessa1211
Wow Respect
wateralien
Getting to the top of HN seems to have crashed this site. Currently.
chancitag
Looks like the personal site is getting overloaded. Archive link here: https://web.archive.org/web/20260629115602/https://haghiri75...
gslin
https://archive.is/q3VyO
zihotki
tldr - Qwen 3.5 9B finetuned on a "simple secret dataset" (500 million tokens of Claude 4.8 Opus, Claude 5 Fable, GPT 5.5 and similar models)
lazerjesus
destillation as a service - coming either from the most or least likely corner of the world. i love it!
misano
slop generated using slops
dingdingdang
OK, I appreciate your excitement and the work you've done here. Also; when presenting a benchmark it -needs- to be done as a comparative measure, i.e. how are those numbers on a vanilla 9b, let alone decent quant of 3.6-35b-a3b? What have you achieved here?