Qwen 27B on a RTX 5090 livestreams attempt at open math problem C(25,15,5)

GuiltyBookkeeper4849 · reddit · 2026-09-20

After a viral 63-hour autonomous run of Qwen 3.8 on an RTX 3090 attempting the Riemann hypothesis, the author launched a new live experiment: Qwen 3.8 27B on a single RTX 5090 tackling the open covering design problem C(25,15,5).

• The best known solution uses 42 groups; the open goal is 41
• The problem was chosen because solutions are trivially verifiable via an independent checker
• The live site exposes the agent's messages, memory, life-loop, code, and runs a public verifier in real time
• The author argues open-source models on consumer hardware can push science forward and that agentic autonomous research should be open to humans and AI agents alike

Original post →

More from Models

• Reddit asks: are Q1 extreme quantizations really that bad, or does architecture matter more? — Felix_455-788 · 2026-09-20
• Dev swaps jev into real browser-use harness: it breaks instantly on real websites — TheZachMueller · 2026-09-20
• $20/mo OpenAI users can't even pick the new "Sol" model — Sauers_ · 2026-09-20
• Ternary 2-bit Bonsai-2-27B GGUF lands on Hugging Face trending — dealignai · 2026-09-20
• Four LLMs play Doom: Jev leads with 5.63 mean kills but 15x higher latency — shniydder · 2026-09-20
• Speculation: China still distilling frontier models implies weights weren't stolen — menhguin · 2026-09-20

Models channel →