{"id":"smaug-audio-decisions-0-7b","slug":"smaug-audio-decisions-0-7b","name":"Smaug-Audio-Decisions-0.7B","orgId":"abacus-ai","family":"smaug-audio-decisions-0-7b","releaseDate":null,"availability":"ga","openWeights":true,"license":"Apache 2.0","paramsTotal":700000000,"paramsSource":"slug","contextWindow":null,"knowledgeCutoff":null,"modelCardUrl":"https://x.com/bindureddy/status/2117659234448154646","hfRepo":null,"modalities":["audio","text"],"aliases":["Smaug-Audio-Decisions-0.7B","smaug audio decisions","smaug-audio-decisions","abacusai/smaug-audio-decisions-0.7b","smaug audio decisions 0.7b","smaug-audio-decisions-0.7b"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-11-03","announced":"2026-11-03","blurb":"Abacus.AI's 744M audio decision model: it answers a typed question about a clip of up to 30 seconds with a calibrated probability per option in one forward pass, on a GPU, a Mac or offline on an iPhone. Fine-tuned from Google's EmbeddingGemma 2; open weights under Apache 2.0, with int8 Core ML builds.","api":"none","openWeights":"yes","params":{"text":"744M (744,371,512 stored parameters, audio tower included)","source":"https://huggingface.co/abacusai/Smaug-Audio-Decisions-0.7B"},"facts":[{"label":"Type","value":"Audio decision model (audio classification): a probability per option, softmax for one answer or sigmoid for select-all; no text generation"},{"label":"Base model","value":"google/embeddinggemma-2 (fine-tuned, with a linear head at a marker token per option)"},{"label":"Input","value":"Audio up to 30 s (16 kHz mono) plus a question with 2 or more options; text and images are accepted, but the card says text decisions are close to chance"},{"label":"Cascade","value":"Designed to sit in front of abacusai/Smaug-Decisions (4B) and hand low-confidence questions to it"},{"label":"Weights","value":"Hugging Face abacusai/Smaug-Audio-Decisions-0.7B: float32 Transformers weights and int8 Core ML models; not gated"},{"label":"License","value":"Apache 2.0"},{"label":"API","value":"No hosted API stated; runs locally, including an iOS app in abacusai/Smaug-Decisions-Mobile"}],"claims":["Lab claim: on top of the leaderboard (Bindu Reddy's post; no leaderboard named).","Lab claim: 53.0 audio skill on Abacus's own 6,671-item Earshot benchmark, against 50.5 for Gemini 3.8 Flash and 43.6 for AudioJev, with 0.086 calibration error.","Lab claim: 9.7 ms per decision on an H100 and about 0.3 s offline on an iPhone 14 Pro Max; as a cascade in front of Smaug-Decisions it matches that model's 60.1 at 130 ms instead of 210 ms."],"sources":[{"kind":"lab","title":"Bindu Reddy (Abacus.AI CEO) announcement on X","url":"https://x.com/bindureddy/status/2117659234448154646","date":"2026-11-03"},{"kind":"repo","title":"abacusai/Smaug-Audio-Decisions-0.7B on Hugging Face","url":"https://huggingface.co/abacusai/Smaug-Audio-Decisions-0.7B"}],"firstBenchmark":null},"organization":{"id":"abacus-ai","name":"Abacus.AI","slug":"abacus-ai","country":null,"url":null,"color":"#11110F","mark":"A"},"price":null,"answer":{"word":"Pending benchmarks.","scope":"Announced Nov 3, 2026; no independent results yet.","lead":"As of Nov 22, 2026, Smaug-Audio-Decisions-0.7B is pending benchmarks on isitsota: announced Nov 3, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}