Close Menu

    Subscribe to Updates

    Get the latest creative news from FooBar about art, design and business.

    What's Hot

    Reform UK Will get Recent $4M Enhance from Tether‑Linked Crypto Investor

    March 5, 2026

    WWE 2K26 Review – A Good Feud That Should Have Main Evented

    March 5, 2026

    PIPOS Pakistan Institute of prosthetic & Orthotics Jobs 2026 2026 Job Advertisement Pakistan

    March 5, 2026
    Facebook X (Twitter) Instagram
    Thursday, March 5
    Trending
    • Reform UK Will get Recent $4M Enhance from Tether‑Linked Crypto Investor
    • WWE 2K26 Review – A Good Feud That Should Have Main Evented
    • PIPOS Pakistan Institute of prosthetic & Orthotics Jobs 2026 2026 Job Advertisement Pakistan
    • S T O P – Neglect All About Manifestation And Numerology…
    • Oak Park Raiders force deciding game with 3-2 OT victory in high school city championship – Winnipeg
    • Sydney Sweeney reveals Cristiano Ronaldo as her sporting inspiration
    • Google Search rolls out Gemini’s Canvas in AI Mode to all US users
    • PSX levels dramatic rebound, surges 5,433 factors
    • Knowledgeable Claims Ripple Is Subsequent to Safe Fed Grasp Account After Kraken Win— Right here’s Why
    • Trump White Home Makes use of Name of Responsibility In New Video Selling Iran Bombings
    Facebook X (Twitter) Instagram Pinterest Vimeo
    The News92The News92
    • Home
    • World
    • National
    • Sports
    • Crypto
    • Travel
    • Lifestyle
    • Jobs
    • Insurance
    • Gaming
    • AI & Tech
    • Health & Fitness
    The News92The News92
    Home - AI & Tech - YuanLab AI Releases Yuan 3.0 Ultra: A Flagship Multimodal MoE Foundation Model, Built for Stronger Intelligence and Unrivaled Efficiency
    AI & Tech

    YuanLab AI Releases Yuan 3.0 Ultra: A Flagship Multimodal MoE Foundation Model, Built for Stronger Intelligence and Unrivaled Efficiency

    Naveed AhmadBy Naveed AhmadMarch 5, 2026No Comments3 Mins Read
    Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
    Share
    Facebook Twitter LinkedIn Pinterest Email


    How can a trillion-parameter Large Language Model achieve state-of-the-art enterprise performance while simultaneously cutting its total parameter count by 33.3% and boosting pre-training efficiency by 49%? Yuan Lab AI releases Yuan3.0 Ultra, an open-source Mixture-of-Experts (MoE) large language model featuring 1T total parameters and 68.8B activated parameters. The model architecture is designed to optimize performance in enterprise-specific tasks while maintaining competitive general-purpose capabilities. Unlike traditional dense models, Yuan3.0 Ultra utilizes sparsity to scale capacity without a linear increase in computational cost.

    Layer-Adaptive Expert Pruning (LAEP)

    The primary innovation in Yuan3.0 Ultra’s training is the Layer-Adaptive Expert Pruning (LAEP) algorithm. While expert pruning is typically applied post-training, LAEP identifies and removes underutilized experts directly during the pre-training stage.

    Research into expert load distribution revealed two distinct phases during pre-training:

    1. Initial Transition Phase: Characterized by high volatility in expert loads inherited from random initialization.
    2. Stable Phase: Expert loads converge, and the relative ranking of experts based on token assignment remains largely fixed.

    Once the stable phase is reached, LAEP applies pruning based on two constraints:

    • Individual Load Constraint (⍺): Targets experts whose token load is significantly lower than the layer average.
    • Cumulative Load Constraint (β): Identifies the subset of experts contributing the least to total token processing.

    By applying LAEP with β=0.1 and varying ⍺, the model was pruned from an initial 1.5T parameters down to 1T parameters. This 33.3% reduction in total parameters preserved the model’s multi-domain performance while significantly lowering memory requirements for deployment. In the 1T configuration, the number of experts per layer was reduced from 64 to a maximum of 48 preserved experts.

    https://github.com/Yuan-lab-LLM/Yuan3.0-Ultra/blob/main/Docs/Yuan3.0_Ultra%20Paper.pdf

    Hardware Efficiency and Expert Rearrangement

    MoE models often suffer from device-level load imbalance when experts are distributed across a computing cluster. To address this, Yuan3.0 Ultra implements an Expert Rearranging algorithm.

    This algorithm ranks experts by token load and uses a greedy strategy to distribute them across GPUs so that the cumulative token variance is minimized.

    MethodTFLOPS per GPU
    Base Model (1515B)62.14
    DeepSeek-V3 Aux Loss80.82
    Yuan3.0 Ultra (LAEP)92.60

    Total pre-training efficiency improved by 49%. This improvement is attributed to two factors:

    • Model Pruning: Contributed 32.4% to the efficiency gain.
    • Expert Rearrangement: Contributed 15.9% to the efficiency gain.

    Mitigating Overthinking with Revised RIRM

    In the reinforcement learning (RL) stage, the model employs a refined Reflection Inhibition Reward Mechanism (RIRM) to prevent excessively long reasoning chains for simple tasks.

    The reward for reflection, $R_{ver}$, is calculated using a threshold-based penalty system:

    • rmin=0: The ideal number of reflection steps for direct responses.
    • rmax=3: The maximum tolerable reflection threshold.

    For correct samples, the reward decreases as reflection steps approach rmax, while incorrect samples that ‘overthink’ (exceeding rmax receive maximum penalties. This mechanism resulted in a 16.33% gain in training accuracy and a 14.38% reduction in output token length.

    https://github.com/Yuan-lab-LLM/Yuan3.0-Ultra/blob/main/Docs/Yuan3.0_Ultra%20Paper.pdf

    Enterprise Benchmark Performance

    Yuan3.0 Ultra was evaluated against several industry models, including GPT-5.2 and Gemini 3.1 Pro, across specialized enterprise benchmarks.

    BenchmarkTask CategoryYuan3.0 Ultra ScoreLeading Competitor Score
    DocmatixMultimodal RAG67.4%48.4% (GPT-5.2)
    ChatRAGText Retrieval (Avg)68.2%53.6% (Kimi K2.5)
    MMTabTable Reasoning62.3%66.2% (Kimi K2.5)
    SummEvalText Summarization62.8%49.9% (Claude Opus 4.6)
    Spider 1.0Text-to-SQL83.9%82.7% (Kimi K2.5)
    BFCL V3Tool Invocation67.8%78.8% (Gemini 3.1 Pro)

    The results indicate that Yuan3.0 Ultra achieves state-of-the-art accuracy in multimodal retrieval (Docmatix) and long-context retrieval (ChatRAG) while maintaining robust performance in structured data processing and tool calling.


    Check out the Paper and Repo. Also, feel free to follow us on Twitter and don’t forget to join our 120k+ ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.




    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    Previous ArticleIceland supermarket drops decade-long trademark dispute with Iceland and offers “rapprochement discount”
    Next Article Pakistan seal quarter-final spot in ITF juniors
    Naveed Ahmad
    • Website
    • Tumblr

    Related Posts

    AI & Tech

    Google Search rolls out Gemini’s Canvas in AI Mode to all US users

    March 5, 2026
    AI & Tech

    MacBook Neo, iPhone 17e, and everything else Apple announced this week

    March 5, 2026
    AI & Tech

    US and EU police shut down LeakBase, a web site accused of sharing stolen passwords and hacking instruments

    March 5, 2026
    Add A Comment
    Leave A Reply Cancel Reply

    Demo
    Top Posts

    How to Get a Bigger Penis – The Stem Cell Secret to Natural Penis Enlargement & A Quiz

    February 22, 20261 Views

    10 Totally different Methods to Safe Your Enterprise Premises

    February 19, 20261 Views

    Oatly loses ‘milk’ branding battle in UK Supreme Courtroom

    February 12, 20261 Views
    Stay In Touch
    • Facebook
    • YouTube
    • TikTok
    • WhatsApp
    • Twitter
    • Instagram
    Latest Reviews

    Subscribe to Updates

    Get the latest tech news from FooBar about tech, design and biz.

    Demo
    Most Popular

    How to Get a Bigger Penis – The Stem Cell Secret to Natural Penis Enlargement & A Quiz

    February 22, 20261 Views

    10 Totally different Methods to Safe Your Enterprise Premises

    February 19, 20261 Views

    Oatly loses ‘milk’ branding battle in UK Supreme Courtroom

    February 12, 20261 Views
    Our Picks

    Reform UK Will get Recent $4M Enhance from Tether‑Linked Crypto Investor

    March 5, 2026

    WWE 2K26 Review – A Good Feud That Should Have Main Evented

    March 5, 2026

    PIPOS Pakistan Institute of prosthetic & Orthotics Jobs 2026 2026 Job Advertisement Pakistan

    March 5, 2026

    Subscribe to Updates

    Get the latest creative news from FooBar about art, design and business.

    Facebook X (Twitter) Instagram Pinterest
    • About Us
    • Contact Us
    • Privacy Policy
    • Terms & Conditions
    • Advertise
    • Disclaimer
    © 2026 TheNews92.com. All Rights Reserved. Unauthorized reproduction or redistribution of content is strictly prohibited.

    Type above and press Enter to search. Press Esc to cancel.