HomeDBRX-Instruct
132B
Model parameter quantity
DataBricks
Affiliated organization
Open Source
License Type
March 26, 2024
Release time
Model Introduction
DBRX-Instruct is an MoE model trained from scratch by DataBricks, utilizing a selection scheme of 16 experts choosing 4, with an active parameter count of 36B. It's pretrained on 12T tokens, supporting a 32K context.
Comprehensive score
Language dialogue
Knowledge reserve
Reasoning association
Mathematical calculation
Code writing
Command following
Swipe left and right to view more
Language comprehension ability
Often makes semantic misjudgments, leading to obvious logical disconnects in responses.
3.8
Knowledge coverage scope
Has significant knowledge blind spots, often showing factual errors and repeating outdated information.
5.9
Reasoning ability
Unable to maintain coherent reasoning chains, often causing inverted causality or miscalculations.
2.6
Model comparison
DBRX-Instruct vs Doubao-Seed-1.6-251015 (Thinking)
The deep reasoning model released by ByteDance, which supports manual switching of deep reasoning, and its performance is significantly improved compared to doubao-1.5.
DBRX-Instruct vs Doubao-Seed-1.6-thinking-250715
The latest version of the seed series model launched by ByteDance, which supports the thinking mode.
Related model
DBRX-Instruct
DBRX-Instruct is an MoE model trained from scratch by DataBricks, utilizing a selection scheme of 16 experts choosing 4, with an active parameter count of 36B. It's pretrained on 12T tokens, supporting a 32K context.
Doubao-Seed-1.6-251015 (Thinking)
The deep reasoning model released by ByteDance, which supports manual switching of deep reasoning, and its performance is significantly improved compared to doubao-1.5.
Doubao-Seed-1.6-thinking-250715
The latest version of the seed series model launched by ByteDance, which supports the thinking mode.
ERNIE 5.0
Baidu’s newly released ERNIE large model, version 5.0.
Spark-X1
The inference model Spark X1 released by iFlytek, on the basis of leading domestic mathematical tasks, benchmarks the performance of general tasks such as inference, text generation, and language understanding against OpenAI o1 and DeepSeek R1.
Relevant documents
Musk Considered Leaving OpenAI to His Kids as Altman Testifies
This morning, OpenAI CEO Sam Altman took the stand to address former co-founder Elon Musk’s lawsuit challenging the company’s corporate structure.When asked about Musk’s claim that other founders “stole a charity” by launching a for-profit subsidiary
Sam Altman Sparks Debate Over AI's Deceleration
Listen onApple PodcastsListen onSpotifyOpenAI CEO Sam Altman recently suggested that it may be time to “pace the rate of AI development” to allow society to “harden around some of these new capability levels.”On the latest episode of TechCrunch’s Equ
Anthropic Opens Doors to EU Cybersecurity Agency as Mythos5 Model Faces Compliance Exam
Artificial intelligence compliance regulations are advancing significantly. Leading AI firm Anthropic has officially granted the European Union's cybersecurity authority access to its Mythos AI model, a pivotal move for this advanced large language m
Lenovo Unveils AI Cutie at MWC 2026: Desktop Robotic Arm Becomes Your New Workplace Assistant
If AI in 2025 is still confined to screen-based chats, 2026 marks the shift toward tangible, desk-integrated intelligence. At MWC 2026 in Barcelona, Lenovo unveiled two groundbreaking AI hardware concepts: AI Workmate (an AI Office Partner) and AI Wo
TikTok Launches Voice Copyright Report Channel as AI Clone Voice Complaints Double
TikTok has introduced a dedicated reporting channel for voice-related intellectual property infringement, alongside enhanced rights protection mechanisms. The platform notes that as AI voice synthesis and imitation technologies become more accessible
