{"schema":"postcutoff/event@1","as_of":"2026-11-19T22:11:00+01:00","url":"https://postcutoff.com/e/2026-10-26-iflytek-sparkaudio-1-0-robot-super-brain-3/","md":"https://postcutoff.com/e/2026-10-26-iflytek-sparkaudio-1-0-robot-super-brain-3/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-10-26-iflytek-sparkaudio-1-0-robot-super-brain-3","date":"2026-10-26","date_precision":"day","short_title":"iFLYTEK details SparkAudio 1.0, a unified speech foundation model for 99 languages trained on Huawei and Hygon chips, and launches Robot Super Brain Platform 3.0","deck":null,"takeaway":null,"category":"model-release","category_label":"Model releases","importance":2,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":[{"n":1,"title":"Pandaily: iFLYTEK details SparkAudio 1.0 speech foundation model and launches Robot Super Brain Platform 3.0 (Oct 26)","url":"https://pandaily.com/iflytek-sparkaudio-1-0-speech-foundation-model-99-languages-robot-super-brain-platform-3-0","type":"press","group":"press","domain":"pandaily.com"}],"official":0,"filed":"2026-10-26","updated":"2026-10-26","orgs":["iFLYTEK"],"title":"iFLYTEK details SparkAudio 1.0, a unified speech foundation model for 99 languages trained on Huawei and Hygon chips, and launches Robot Super Brain Platform 3.0","summary":"At its 2026 Global 1024 Developer Festival, iFLYTEK presented SparkAudio 1.0, a speech foundation model that merges speech recognition, translation and semantic understanding (0.65B audio encoder + 30B-A3B MoE language model, 13M hours of audio, 99 languages and 202 dialects, trained on Huawei Ascend 910B and Hygon BW1000 accelerators), and Robot Super Brain Platform 3.0 for embodied multimodal interaction. Reported by Pandaily on Oct 26, 2026; performance figures are iFLYTEK's own.","key_facts":["SparkAudio 1.0 architecture (technical report, via Pandaily): 0.65B-parameter dense audio encoder + 30B-A3B mixture-of-experts language model; trained on 13M hours of audio plus text","Coverage: speech recognition in 99 languages and 202 dialects; also real-time speaker separation, environmental-sound and audio understanding, emotion recognition","Compute: trained on domestic accelerators including Huawei Ascend 910B and Hygon BW1000","Chairman Liu Qingfeng's claimed gains: semi-supervised audio decoder +20% accuracy; dynamic segmented end-to-end modelling −15% response time; dual-paradigm collaborative decoding −60% cost (company figures; the report says instruction following and complex audio reasoning still need work)","Shipping in iFLYTEK Input Method 16.0, iFLYTEK Hearing, smart notebooks and translators; an AI recording ring said to transcribe two hours of audio in two minutes","Robot Super Brain Platform 3.0: Spark super-brain board, multimodal interaction module, embodied AI dev kit, and a device–cloud interaction harness so a robot can keep talking while doing tasks; demoed in an unstaffed 24-hour robot-run store","Date: Pandaily's report is dated Oct 26, 2026; the exact keynote day was not confirmed"],"key_numbers":[],"tags":["speech","audio","china","robotics","domestic-chips","ascend"],"science":null,"body_md":"## What happened\n\niFLYTEK, China's largest speech-technology company, used its annual developer festival to lay out a path from voice interaction to\nmultimodal and then embodied interaction for robots. Pandaily drew on Sina Tech, IT Home and the transcript of Liu Qingfeng's keynote.\nWe did not find the SparkAudio technical report itself; the architecture numbers come from Pandaily's reading of it.\n\n## Why it matters\n\nA large speech model trained entirely on Chinese accelerators (Ascend and Hygon) is one more data point that export controls have not\nstopped domestic training of production models, at least at this scale.","disputed":[],"related":[{"id":"2026-10-23-block-dessa-speech-only-full-duplex-model","url":"https://postcutoff.com/e/2026-10-23-block-dessa-speech-only-full-duplex-model/","date":"2026-10-23","date_precision":"day","short_title":"Block unveils Dessa, a full-duplex voice model trained from scratch only on recorded human conversation, with no text and no pretrained model","deck":null,"takeaway":"Most voice assistants of 2026 convert speech to text or are built on text-trained models.","category":"research","category_label":"Research","importance":3,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":7,"official":4,"filed":"2026-10-24","updated":"2026-10-26","orgs":["Block"]}],"people":[{"id":"liu-qingfeng","name":"Liu Qingfeng","url":"https://postcutoff.com/person/liu-qingfeng/"}],"posts":[],"videos":[],"models":[],"changes":[{"date":"2026-10-26","type":"filed","text":"Created. Technical report not located"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-10-26","run":null,"sources_read":null,"updated":"2026-10-26","human_review":null,"version":{"date":"2026-10-27"}},"gaps":[{"model_id":"gpt-6-2-astra","name":"GPT-6.2 Astra","cutoff":"2026-05","days_after":148,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":118,"in_training_data":false},{"model_id":"gemini-4-hydrogen","name":"Gemini 4 Hydrogen","cutoff":"2026-03","days_after":209,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":148,"in_training_data":false}],"short_url":null}