Model Detail
MOSS-VL-Base-0708
—MOSS-VL-Base-0708 is a code generation model with 5.7B parameters released by OpenMOSS-Team. The model is registered under the video-text-to-text pipeline tag on Hugging Face, distributed under the permissive apache-2.0 license.
MOSS-VL-Base-0708 ships with 5.7B parameters. Total weight footprint is approximately 11.3 GB, which is the relevant figure when planning local-inference VRAM. The apache-2.0 license is permissive, allowing commercial deployment and derivative work without per-seat fees, though attribution requirements still apply.
MOSS-VL-Base-0708 is best fit for code completion, repository-scale Q&A, and pair-programming integrations. It is a less obvious choice for one-shot generation of security-critical code without review. Treat this as a starting matrix rather than a benchmark verdict — the right deployment usually depends on the specific evaluation suite that mirrors your workload.
MOSS: A Production Framework for Monitoring and Diagnosing Semantic Failures in Search and Recommendation Systems
arXiv:2610.26902v1 Announce Type: cross Abstract: Detecting semantic failures in production recommendation and search systems is hard: behavioral metrics lag by days, manual review does not scale, and most LLM evaluation frameworks are designed for offline, fixed datasets rather than continuous oper
MOSS-Audio Technical Report
arXiv:2606.01802v2 Announce Type: replace-cross Abstract: MOSS-Audio is a unified audio-language model for speech, environmental sound, and music understanding, supporting audio captioning, time-aware question answering, timestamped transcription, and audio-grounded reasoning. MOSS-Audio couples a d
Rel-MOSS: Towards Imbalanced Relational Deep Learning on Relational Databases
arXiv:2603.07916v2 Announce Type: replace Abstract: In recent advances, to enable a fully data-driven learning paradigm on relational databases (RDB), relational deep learning (RDL) is proposed to structure the RDB as a heterogeneous entity graph and adopt the graph neural network (GNN) as the predi
MOSS: Self-Evolution through Source-Level Rewriting in Autonomous Agent Systems
arXiv:2605.22794v2 Announce Type: replace Abstract: Autonomous agentic systems are largely static after deployment: they do not learn from user interactions, and recurring failures persist until the next human-driven update ships a fix. Self-evolving agents have emerged in response, but all confine
