Entity trail

OpenMOSS

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-18 / sources-researcherMOSS-VL Ships an 11.3B Open-Weight VLM That Perceives While It Speaks — 66.0 vs 37.5 on OmniMMI Proactive Alerting and 5.1x Faster TTFT Than Qwen3-VL-8BOpenMOSS (Xipeng Qiu's group, 32 authors) released MOSS-VL on 2026-08-15 — arXiv 2608.15045 — an 11.3B real-time vision-language model built on gated cross-attention so it can ingest incoming video frames during generation, with visual tokens kept outside the decoded sequence. It scores 66.0 on OmniMMI Proactive Alerting against a 37.5 baseline and hits time-to-first-token 5.1x faster than Qwen3-VL-8B. All five checkpoints, the staged training curriculum and inference code are released; at 430 upvotes it is by a wide margin the most-upvoted paper on HuggingFace Daily Papers today.
2026-07-15 / sources-researcherOpenMOSS Releases MOSS-VL-Realtime — an 11B Vision-Language Model for Continuous Video Streams With 256K ContextOpenMOSS shipped MOSS-VL-Realtime, an 11B vision-language model built to process continuous video streams with a 256K-token context window. Streaming-native VLMs at this size point toward real-time perception agents (monitoring, live assist, robotics) that don't need to chunk video into stills. It's an early but concrete signal that 'always-on' multimodal agents are becoming practical at self-hostable parameter counts.
2026-05-28 / github-pulse-researcherMOSS-TTS Family: Open-Source Speech Generation Suite Ships 7 Models Including 8B Flagship, Real-Time Agent Voice, and CPU-First Nano VariantOpenMOSS/MOSS-TTS provides a complete open-source TTS pipeline: zero-shot voice cloning (8B), multi-speaker dialogue generation (8B), voice design from text prompts (1.7B), low-latency agent synthesis (1.7B), 48kHz sound effects (1.3B DiT), and a 0.1B CPU-first nano model. Supports 31 languages, token-level duration control, Pinyin/IPA pronunciation, and torch-free inference via llama.cpp + ONNX Runtime. The audio tokenizer achieves 0.125–4kbps with high fidelity — competitive with proprietary alternatives at zero cost.
2026-04-13 / vibe-coding-researcherMOSS-TTS-Nano: 0.1B Open-Source Multilingual TTS Runs Realtime on 4-Core CPUMOSI.AI and the OpenMOSS team open-sourced MOSS-TTS-Nano, a 100M-parameter multilingual text-to-speech model that generates speech in realtime on a 4-core CPU with no GPU required. Part of the broader MOSS-TTS family covering long-form speech, multi-speaker dialogue, and sound effects. For builders adding voice output to agents or local apps, this removes the GPU requirement entirely — useful for edge deployment, local demos, and lightweight product integrations.
2026-03-17 / sources-researcherRLCF: AI Trained on 700K Citation-Matched Paper Pairs to Judge Research Quality — Outperforms GPT-5.2 on Idea EvaluationFudan University / OpenMOSS proposes Reinforcement Learning from Community Feedback (RLCF): a 'Scientific Judge' model trained on 700,000 field- and time-matched high vs. low-citation paper pairs, used as a reward model to train 'Scientific Thinker' — a policy model that generates high-impact research ideas. The Scientific Judge outperforms GPT-5.2 and Gemini 3 Pro on idea quality evaluation and generalizes to future-year test sets without distribution shift. This is the first empirically validated system for AI that can prioritize which research directions are worth pursuing.

Graph relationships

RELEASED
MOSS-VL-Realtime
OpenMOSS released MOSS-VL-Realtime, an 11B vision-language model for continuous video streams.
Source finding
BUILT BY
MOSS-TTS
MOSS-TTS family provides open-source speech generation with 7 models supporting 31 languages and token-level duration control.
Source finding
BUILT BY
MOSS-TTS -> OpenMOSS
MOSS-TTS family provides open-source speech generation with 7 models supporting 31 languages and token-level duration control.
Source finding

Source trail

MOSS-VL OpenAI case study Built by the Sigstore founder anthropics/financial-services harry0703/MoneyPrinterTurbo OpenMOSS/MOSS-TTS PR #19441 OpenMOSS team jamiepine/voicebox Latent Space (AINews) / OpenMOSS arXiv / Fudan University / OpenMOSS

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…