PulseAugur
EN
LIVE 19:55:49

DeepSeek releases new models outperforming Nemotron3 Ultra

By PulseAugur Editorial · [2 sources] ·

DeepSeek has released two new models, DeepSeekv4 Pro 0813 1.5T and DeepSeekv4 Flash 0731, which reportedly outperform NVIDIA's Nemotron3 Ultra on agentic tasks. The DeepSeekv4 Flash model achieves this with significantly fewer parameters. The analysis also suggests that Mistral AI's pretraining efforts for Nemotron were unsuccessful, and the committee-based development structure of Nemotron is hindering its progress. AI

IMPACT DeepSeek's new models may set a new benchmark for agentic tasks, potentially pressuring competitors like NVIDIA and Mistral AI.

RANK_REASON The cluster reports on the release of new frontier models by DeepSeek, with comparative performance claims against competitor models. [lever_c_demoted from frontier_release: ic=2 ai=1.0]

Read on X — SemiAnalysis →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

DeepSeek releases new models outperforming Nemotron3 Ultra

COVERAGE [2]

  1. X — SemiAnalysis TIER_1 English(EN) · SemiAnalysis_ ·

    @deepseek_ai Furthermore, Mistral fumbled Nemotron's pretraining. NVIDIA's single-committee structure means that one member's failure drags everyone down with i

    @deepseek_ai Furthermore, Mistral fumbled Nemotron's pretraining. NVIDIA's single-committee structure means that one member's failure drags everyone down with it, and the rest of the committee is upset. 2/3🧵

  2. X — SemiAnalysis TIER_1 English(EN) · SemiAnalysis_ ·

    Congrats to @deepseek_ai for their release of DeepSeekv4 Pro 0813 1.5T🔥 It massively beats Nemotron3 Ultra on agentic tasks.

    Congrats to @deepseek_ai for their release of DeepSeekv4 Pro 0813 1.5T🔥 It massively beats Nemotron3 Ultra on agentic tasks. This is in addition to DeepSeekv4 Flash 0731 beating Nemotron3 Ultra massively too, while having 4.2x fewer active parameters and close to 2x fewer total …

RELATED ENTITIES

RELATED TOPICS