1. X
  2. TNG Technology Consulting GmbH
Log inSign up
TNG Technology Consulting GmbH
1,658 posts
Image
user avatar
TNG Technology Consulting GmbH
@tngtech
TNG, aka "The Nerd Group", is a consulting partnership focused on high end information technology, particularly AI. 933 employees, 99.9% academics, ~53% PhDs.
Unterföhring, Deutschland
tngtech.com/en
Joined December 2010
176
Following
2,197
Followers
RepliesRepliesArticlesArticlesMediaMedia

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • Pinned
    user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Jul 2, 2025
    Today we release DeepSeek-TNG R1T2 Chimera. This new Chimera is a Tri-Mind Assembly-of-Experts model with three parents, namely R1-0528, R1 and V3-0324. R1T2 operates at a sweet spot in intelligence vs. output token length. It appears to be... * about 20% faster than R1, and
    R1T2-Chimera: Intelligence Score vs. Inference Cost
    127K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Apr 27, 2025
    Today we release DeepSeek-R1T-Chimera, an open weights model adding R1 reasoning to @deepseek_ai V3-0324 with a novel construction method. In benchmarks, it appears to be as smart as R1 but much faster, using 40% fewer output tokens. The Chimera is a child LLM, using V3s
    R1T-Chimera
    81K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Mar 17, 2025
    Inference test - @AMD MI300X beats @nvidia H100 We got access to 8x MI300X - 192 GB each and therefore 1.5 terabyte of vRAM in total - provided by @Dell Running @deepseek_ai R1, we compared inference metrics against 8x H100 - 94 GB each Here are the results 🚀 1/n
    MI300X vs H100
    15K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Sep 6, 2024
    Unlock the math behind Transformers! Grant Sanderson (@3Blue1Brown) shared insights on attention blocks, parallel processing & more at our Big Techday 24. Watch the full talk here: youtu.be/rZTV_1TcVyY #tngbtd #Transformers
    Image
    Big Techday 24: Visualizing Transformers - Grant Sanderson (@3blue1brown)
    From youtube.com
    26K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Apr 29, 2025
    Oh man, lucky day 😅 R1T-Chimera is ranked the #2 trending model on @openrouter Sure, "trending" is a temporary attention metric, not reflecting total usage. And the world spins fast: everybody talks about Qwen3 now. Still a nice screenshot with @GoogleAI, @Microsoft,
    R1T Chimera #2 trending model
    11K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Jul 8, 2025
    Image
    Image
    9.1K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Jun 1, 2025
    More evidence for the effectiveness of the Chimera construction method: Taking DeepSeek's R1-0528 release, we started benchmarking new Chimera variants on AIME-24 and SimpleQA. R1-0528 significantly improves AIME performance from 79.8 to 91.4 while doubling the amount of output
    Image
    Image
    9.1K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Jul 9, 2025
    DeepSeek-TNG-R1T2-Chimera is currently the #1 trending model on @openrouter The platform which @karpathy called the "transfer switch of AI". Our Assembly-of-Experts method pushes the Pareto frontier between model intelligence and inference cost. Thanks again to the Open
    Image
    Image
    Image
    user avatar
    Deedy
    Menlo Ventures
    @deedydas
    Jun 25, 2025
    Karpathy at YC startup school calls this the transfer switch of AI. It's hard to keep up with all the LLMs out there. OpenRouter is the go-to place for 2.5M developers to choose from 400+ models with one API. They're serving 100T tokens/yr! Excited to back this special company!
    12K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Apr 27, 2025
    Replying to @tngtech
    Link to model weights on @huggingface: huggingface.co/tngtech/DeepSe… For people without the infrastructure to run the 671B model, we can provide test access to our R1T cluster (DM us)
    Image
    tngtech/DeepSeek-R1T-Chimera · Hugging Face
    From huggingface.co
    4.1K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    May 7, 2025
    Eight new @AMD MI325X GPUs joined our compute cluster of @nvidia H100s. The new @Supermicro_SMCI server is an AI machine with spectacular 2 Terabytes of total GPU memory in one ~10kW node. ROCm worked right away with full VRAM and GPU utilization, allowing new types of
    Image
    3.3K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Jul 3, 2025
    R1T2 Chimera is available on @chutes_ai Thanks to @jon_durbin and team for making it available within 24 hours!
    Image
    5.5K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Jun 16, 2025
    We post our new paper "Assembly of Experts: Linear-time construction of the Chimera LLM variants with emergent and adaptable behaviors" on @huggingface, while waiting for @arxiv. We explain how we constructed the 671B R1T Chimera child model from the great @deepseek_ai V3-0324
    Image
    13K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Aug 7, 2025
    Big update: DeepSeek-TNG-R1T2-Chimera now fully supports tool calling It's already live on @openrouter and @chutes_ai. How does it work? Here's the technical detail: We provide an extended chat template together with a custom Chimera tool-call-parser. Good news: vLLM can be
    Image
    Image
    5.6K
  • user avatar
    TNG Technology Consulting GmbH
    @tngtech
    Apr 27, 2025
    You can now vibe check DeepSeek-R1T-Chimera yourself Thanks to @openrouter for hosting our new model
    Image
    4.3K
Advertisement
Advertisement