{"id":"avalon-m1-mini","slug":"avalon-m1-mini","name":"Avalon-M1-Mini","orgId":"general-reasoning","family":"avalon-m1-mini","releaseDate":null,"availability":"announced","openWeights":false,"license":null,"paramsTotal":null,"paramsSource":null,"contextWindow":null,"knowledgeCutoff":null,"modelCardUrl":"https://x.com/GenReasoning/status/2113590771722952908","hfRepo":null,"modalities":["text"],"aliases":["Avalon-M1-Mini","avalon-m1-mini","avalon m1 mini","avalon-m1 mini","general reasoning avalon-m1-mini"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-10-23","announced":"2026-10-23","blurb":"General Reasoning's research model trained to work as an agent organisation: copies of one policy act as a CEO and hired specialists sharing a message board and workspace, rewarded as a team (reinforcement learning from collective rewards, RLCR). Post-trained from NVIDIA's Nemotron 3 Super base. Training is still running; results so far are a preview, and weights are planned but not out.","api":"none","openWeights":"announced","params":{"text":"120B total / 12B active","source":"General Reasoning's Avalon-M1 release page"},"facts":[{"label":"Base model","value":"NVIDIA Nemotron 3 Super pretrained base (120B-A12B)"},{"label":"Recipe","value":"Reinforcement learning from collective rewards (RLCR): every agent in the organisation gets the same collective advantage"},{"label":"Training data","value":"Small SFT set of 38,374 examples, then RL over 147 environments with more than 2 million tasks"},{"label":"Status","value":"Still training; the technical report is marked \"under revision\""},{"label":"Weights","value":"Not released; the lab's roadmap plans to release the model weights, harness and open environments"},{"label":"API","value":"None announced"},{"label":"Paper","value":"Technical report on gr.inc; final version planned for arXiv"}],"claims":["Lab claim: in the lab's organisational evaluations, Avalon-M1-Mini outperforms Nemotron 3 Super, post-trained from the same base without on-policy distillation, on every benchmark shown.","Lab claim: against Inkling Small, comparing each model's better configuration, Avalon-M1-Mini wins 6 of 18 benchmarks and ties 3 (all run on the lab's own infrastructure).","Lab claim: working as an organisation instead of alone raises IMO-ProofBench-Advanced from 29.8% to 37.5% and FrontierScience from 55.9% to 63.5%."],"sources":[{"kind":"lab","title":"General Reasoning announcement on X","url":"https://x.com/GenReasoning/status/2113590771722952908","date":"2026-10-23"},{"kind":"lab","title":"Avalon-M1 release page","url":"https://www.gr.inc/releases/avalon-m1","date":"2026-10-23"},{"kind":"paper","title":"Avalon-M1: Reinforcement Learning from Collective Rewards (technical report)","url":"https://www.gr.inc/avalonm1.pdf","date":"2026-10-23"},{"kind":"lab","title":"Ross Taylor (General Reasoning) post on the base model","url":"https://x.com/rosstaylor90/status/2113616941646573716","date":"2026-10-23"}],"firstBenchmark":null},"organization":{"id":"general-reasoning","name":"General Reasoning","slug":"general-reasoning","country":null,"url":"https://www.gr.inc","color":"#11110F","mark":"G"},"price":null,"answer":{"word":"Pending benchmarks.","scope":"Announced Oct 23, 2026; no independent results yet.","lead":"As of Oct 27, 2026, Avalon-M1-Mini is pending benchmarks on isitsota: announced Oct 23, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}