AI Briefing
KOSign in

Nace releases Drex Max v1.1, ranking #1 on Artificial Analysis for multimodal decision models

·2026.11.09 02:54

Key point

The model features a 256k context window and runs on local hardware via the Kapitsa inference engine.

Details

  • Nace released Drex Max v1.1, which ranks #1 on the Artificial Analysis index for multimodal decision models.
  • The model is 8x cheaper than the next best model in the category and achieves both SOTA and Pareto frontier performance.
  • Drex Max v1.1 offers a 256k context window, significantly larger than Glide1.2 (64k) and Mammoth (16k), while matching pplxdecider.
  • Open weights for Drex Max v1 are available on Hugging Face under the Open RAIL-M license, with v1.1 weights upcoming after local inference optimization.
  • A technical paper titled "Towards Foundation Decision Models: Drex" details the training process and is available on arXiv and Hugging Face.
  • The model can run on local hardware, including Macs, using the Kapitsa inference engine.
open-sourcelocal-inferenceartificial-analysisdecision-modelsnacedrex-max
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

RedditNace AI Releases Drex Decision Models with 183ms Latency and Top 3 OpenRouter RankingRedditNace AI Research Lab publishes 50-page technical report on training SOTA decision modelsKakaoKakao Releases Open-Source 3B Multimodal Decision-Making Model 'Kanana-2-v-decide'PyTorchKRKOMOS Achieves SRCC of 0.900 in Korean Speech MOS PredictionPyTorchKRjev-chat: Uses Jev model to assess counterpart intent and inspect LLM reply candidatesHacker NewsVideo Rebirth Releases HyperFlow H3 Flash, a 13.7B Compressed Video-Audio Diffusion Model