Supra2-Medium: a 25M model trained from scratch on two RTX 5060s beats its 50M predecessor
LH-Tech_AI · reddit · 2026-08-20
- SupraLabs releases Supra2-Medium-Base, a 25M-parameter model built on the Qwen3 architecture and trained entirely from scratch on consumer hardware: an RTX 5060 Ti 16GB plus an RTX 5060 8GB.
- Benchmark results show it competing heavily with the team's previous Supra-50M-Base, which is twice its size.
- This is a base model only; an instruction-tuned version may come later. Model available at huggingface.co/SupraLabs/Supra2-Medium-Base.
- The team teases Supra3 with four tiers: Flash-Lite 25M, Flash 50M, Pro 75M, and Ultra 100M.
More from Models
- Model at capacity for a straight week drives frustrated dev back to Opus — joonasvirtanen · 2026-10-16
- Decision models match LLMs on routing and safety — and are 7-8x faster — randal_olson · 2026-10-16
- Xiaomi MiMo V2.6 Pro takes #1 open-model spot on Agent Arena at $0.11/task — arena · 2026-10-16
- ChatGPT's 'best AI video tool' answer points to Runway integration, sparking ad doubts — tlakomy · 2026-10-16
- Microsoft's MAI finishes in-house 3T-parameter pretrained base models, post-training results coming — soumitrashukla9 · 2026-10-16
- Hands-on with Gemini 4 Argon: faster than Opus 5.5 but hallucinates more and costs more — prasenx · 2026-10-16