Databricks announced DBRX on March 27, 2024, describing it as a general-purpose, open-weight large language model for organizations building and serving customized AI models. Its defining feature is a mixture-of-experts design: DBRX has 132 billion total parameters, but routes each input through four of its 16 experts, activating 36 billion parameters. Databricks reported strong launch-era benchmark results and several access routes; those figures and availability claims should be read in their 2024 context, not as current rankings or guarantees.
What Databricks announced
In its March 27, 2024 announcement, Databricks introduced DBRX as a general-purpose decoder-only transformer. The company positioned it for businesses that want to build and serve customized models, and said it outperformed established open models on selected evaluations. Those performance comparisons were Databricks’ launch claims, not a permanent or independent ranking.
Databricks co-founder and CEO Ali Ghodsi described the launch rationale as benchmark performance, results against GPT-3.5, and the speed and serving-cost potential of mixture-of-experts architecture. His statement was the company’s characterization of DBRX at launch, rather than an independent assessment.
How DBRX’s mixture-of-experts design works
A dense model uses its parameter set for each input. DBRX instead has 16 expert subnetworks and selects four for a given input. Databricks reported 132 billion total parameters and 36 billion active parameters per input. The distinction matters: total parameters describe the model’s overall learned capacity, while active parameters help describe the computation used for an individual input. It does not, by itself, guarantee lower latency or cost in every deployment.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →#1 Best Overall
- Built for Local AI Development: AMD Ryzen AI Halo is designed for local AI development and inference, featuring 128GB unified memory and support for up to 200B parameter models to build and run intensive AI workloads locally.
- 128GB Unified Memory: Features 128GB LPDDR5x unified memory at 8000 MT/s with 256 GB/s memory bandwidth, providing a shared memory pool across the CPU, GPU, and NPU to support larger AI models.
- AMD Ryzen AI Max+ 395 Processor: Features 16 cores, 32 threads, and Zen 5 architecture, paired with AMD Radeon 8060S integrated graphics featuring 40 RDNA 3.5 compute units and an AMD XDNA 2 NPU with up to 50 TOPS.
- Linux AI Developer Platform: Purpose-built for Linux-based AI development with full AMD ROCm software support and preloaded tools, models, and workflows optimized for local AI development.
- Compact, Connected Design: Includes a 2TB M.2 SSD, 10GbE LAN, Wi-Fi 7, Bluetooth 5.4, USB-C connectivity, and HDMI 2.1b.
Databricks’ AI Research Team described DBRX as a next-token-prediction model with rotary position encodings, gated linear units, grouped-query attention, and the GPT-4 tokenizer as implemented in tiktoken. The team said DBRX was pretrained on 12 trillion tokens of curated text and code and supports a maximum context length of 32K tokens.
The company also described its training process: curriculum learning and internal data, training, and experiment-management tools; a run using 3,072 NVIDIA H100 GPUs connected by 3.2 Tbps InfiniBand; and three months encompassing pretraining, post-training, evaluation, red-teaming, and refinement. These are Databricks’ reported details about its own training effort.
What the reported benchmark results show
In its March 2024 research post, Databricks reported the following DBRX Instruct results alongside Mixtral Instruct on two composite evaluations:
| Evaluation | DBRX Instruct | Mixtral Instruct | What the comparison represents |
|---|---|---|---|
| Hugging Face Open LLM Leaderboard composite | 74.5% | 72.7% | Scores reported in Databricks’ March 2024 blog; not a current leaderboard ranking. |
| Databricks Model Gauntlet | 66.8% | 60.7% | Scores reported in Databricks’ March 2024 blog on its evaluation suite. |
The same post reported 70.1% on HumanEval and 66.9% on GSM8k for DBRX Instruct. Databricks noted that results came from a mix of its own measurements and figures reported by the leaderboard or papers, and that a newer evaluation harness changed GSM8k results. A score should therefore be understood in relation to its benchmark version, setup, and source—not as a universal measure of model quality.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesRank #2
- EVOLUTION AMD RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 64GB pool, which is perfect for running LLMs such as Deepseek 32B, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 4% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
For a practical comparison with Mixtral, Llama 2, or another model, look beyond a single composite score. Check the task and evaluation version, total versus active parameters, throughput on comparable hardware and precision, weight-license terms, and deployment route, region, and cost. The launch figures do not settle those questions for a specific workload.
What Databricks said about speed
Databricks reported that DBRX inference could be up to twice as fast as LLaMA 2 70B and that its Model Serving platform could deliver up to 150 tokens per second per user. These are company-reported, launch-era figures under specified optimized serving conditions, not expected speeds on arbitrary hardware.
The detailed comparison depended on the infrastructure, TensorRT-LLM, precision settings, prompt and response lengths, and concurrency used in the test. Throughput can change with serving configuration and workload; compare models only when those conditions are sufficiently alike.
Was DBRX open source?
Databricks used “open” in its launch materials and said DBRX Base and DBRX Instruct weights were available on Hugging Face under an open license, for research and commercial use. The supported description is open-weight: the cited launch materials establish access to model weights, but do not establish that the training corpus or the complete training pipeline was released.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #3
- Intel Core Ultra 9 285 Processor: Newly developed cores deliver ultra-smooth and responsive gameplay. AI accelerators prepare users for the next era of gaming on an AI PC.
- Simplistic Design: Enjoy the latest generation of Windows 11 Home for your everyday needs. *MSI recommends Windows 11 Pro for business use.
- NVIDIA GeForce RTX 5070 Ti GPU
- Cool While Gaming: In conjunction with an RGB CPU Air Cooler, the Aegis RS features four system cooling fans; three in the front and one in the rear to pull in cool air and push heat out of the PC.
- Turn on the Bright Lights: With the built-in RGB lighting, take your gaming experience to the next level by pressing the MSI LED button to cycle through lighting options. Customize lighting even further with MSI Center software.
The launch sources do not provide enough license detail to determine every condition or use case. Review the operative license accompanying the weights before making a commercial or redistribution decision; the phrase “open” alone is not a substitute for those terms.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How DBRX could be accessed at launch—and what to verify now
In March 2024, Databricks named GitHub and Hugging Face for model access, and listed Databricks, AWS, Google Cloud, and Azure Databricks as deployment routes. Its research post also described API access, pay-as-you-go use, provisioned throughput, and private hosting through Databricks. These are historical launch-era options, not confirmation that every route remains available.
Before planning a deployment, check current provider documentation for DBRX support, endpoint status, regional coverage, pricing, and applicable license terms. The Databricks documentation page listing supported Foundation Model APIs, inspected September 28, 2026, does not establish DBRX support in the material available at that time. A 2024 announcement cannot establish present-day service availability or price.
Quick Recap
- For self-managed use, confirm the current weight location, license, infrastructure requirements, and serving software.
- For a hosted deployment, confirm that the provider currently offers the model in the region and service tier you need.
- For a performance estimate, reproduce or closely match the benchmark’s hardware, precision, prompt/output lengths, and concurrency.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →
