DBRX Instruct

+ contribute

DBRX Instruct is an open-weights language model from Databricks, released March 2024. Its published size is 132B parameters (36B active per token) and a 32k-token context window. Licence: databricks-open-model.

Catalogue identity notes

Comparison model from the Epoch AI internal-evaluation CSVs (Epoch release date 2024-03-27). DBRX Instruct (March 2024). 132B total, 36B active parameters and the Databricks Open Model License per the ollama library page/licence layer; the original Hugging Face repo databricks/dbrx-instruct is no longer publicly accessible (HTTP 401, observed 2026-09-28).

45.8±0.7#428 of 719 LLMsfrom 13 benchmark results across 9 sources

$ how it compares

The LLMs ranked either side of DBRX Instruct, out of 828 with a score, then more from Databricks.

#	model	score	where	RAM	released
507	Qwen2.5-Math 1.5B Instruct Alibaba	46±2	local	~2.4GB	2024-09
508	Yi 1.5 34B Chat 01.AI	46±1	local	~22GB	2024-05
509	DBRX Instruct Databricks	46±1	local	~85GB	2024-03
510	Granite 34B Code Instruct (8K) IBM	46±1	local	~22GB	2024-04
511	WizardCoder Python 34B V1.0 WizardLM	46±1	local	~23GB	2023-08
	more from Databricks
805	Dolly v2 12B Databricks	17±2	comparison only		2023-04

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #286 of 300 * · 0.8% of its score · 1 test

Trails Granite 4.0 H 350M, ChatGLM3-6B, Qwen3 0.6B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
BFCL v3	39.2% ✎	▁▁▁▁▃▃▅█▅▄▁▁ beats 13/114	medium · avg 52%	▇ 0.8%	blunt · at its level

coding▁▁▁▃▆██▇▆▄▃▁ #285 of 458 · 61% of its score · 3 tests

Beats Qwen3-VL 32B Instruct, Solar Open 100B (102B-A12B), GLM-4.7-Flash here, though they rank higher overall. Trails DeepSeek LLM 67B Chat, WizardLM 70B V1.0, Yi 1.5 9B Chat here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
HumanEval	75.0% ✎	▁·▁▁▃▄▆▅█▃▁▁ beats 71/99	easy · avg 79%	▇▇▇▇▇▇▇▇▇▇▇▇▇ 31%	sharp · at its level
HumanEval+	70.1% ✎	▁·▁▁▁▃▄▄▄█▃▁ beats 70/99	easy · avg 74%	▇▇▇▇▇▇▇▇▇▇ 25%	sharp · at its level
MBPP+	55.8% ✎	▁···▁▁▂▅▇█▆▂ beats 30/75	easy · avg 68%	▇▇ 4.5%	at its level

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #308 of 393 · 22% of its score · 4 tests

Beats Gemma 4 E4B, OpenThinker3 7B, Gemma 4 E2B here, though they rank higher overall. Trails LFM2 350M, Gemma 3 1B, Flan-T5 XXL here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
AlpacaEval 2 length-controlled	25.4% ✎	▁▂▂▃▆▆█▆▄▄▁▁ beats 66/122	medium · avg 39%	▇▇▇▇▇▇ 14%	at its level
Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024)	1101.0 ✎	▂▁▂▁▃▇█▆▇▄▂▂ beats 45/95	avg 1201	▇▇ 3.8%	blunt
LMArena (text)	1195.5 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 66/379	avg 1268	▇▇ 3.8%	blunt
IFEval (aggregation unspecified)	52.1% ✎	▁▁▃▁▂▄▄▆█▃▂▁ beats 13/141	easy · avg 73%	▇ 0.7%	at its level · small run

math▁▂▃▅▆▅▄█▄▁▁▁ #346 of 398 * · 5.5% of its score · 1 test

Trails Qwen3 0.6B, Gemma 7B (base), Qwen2.5 1.5B Instruct here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
MATH level 5	11.7% ✎	▂▂▅▆▄█▆▃▅▂▃▄ beats 9/92	hard · avg 32%	▇▇ 5.5%	sharp

other tests4 not in a task · 10% of its score

test	result	vs the others in its run	difficulty	counts	why
WildBench v2 WB-Score (task macro)	32.5989 WB-Score ✎	▁·▁▁·▃█▄▇▆▅▆ beats 20/52	avg 45.199 WB-Score	▇▇ 4.3%	sharp
GPQA Diamond	32.9% ✎	▁▁▁▁▃▄▆▄▅▇█▅ beats 21/208	hard · avg 40%	▇ 3.0%	sharp
BIG-Bench Hard	48.5% ✎	▁▁▁▁▂▄▅▅█▃▄▁ beats 42/129	medium · avg 67%	▇ 2.8%	sharp · at its level · small run
Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate)	91.7% ✎	▁▁▂▃▄▄█▆▃▂▂▁ beats 34/140	very easy · avg 94%	▇ 0.3%	blunt · maxed out · noisy

$ where to run it

Ollama · recommended

version

Install Ollama
download it (macOS / Windows) ↗
Linux
curl -fsSL https://ollama.com/install.sh | sh
Run DBRX Instruct · needs ~79.7 GB · from ollama.com
chat in the terminal
ollama run dbrx:132b

$ how the score was worked out · EvalPlus / EvalPerf published… 61%, AlpacaEval 1.0 and 2.0 leader… 14%, Epoch AI — internal evaluatio… 8% and 6 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

source	results	share of the score	licence
EvalPlus / EvalPerf published evaluation runs	3	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 61%	Apache-2.0
AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)	1	▇▇▇▇ 14%	Apache-2.0
Epoch AI — internal evaluations	2	▇▇▇ 8%	CC BY 4.0
WildBench v2 leaderboard data (Ai2)	1	▇ 4%	Apache-2.0
Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)	1	▇ 4%	CC-BY-4.0
LMArena leaderboard dataset	1	▇ 4%	CC BY 4.0
HuggingFaceH4 zephyr-orpo-141b-A35b-v0.1 model card	2	▇ 4%	Apache-2.0
Berkeley Function Calling Leaderboard result archive	1	▇ 1%	Apache-2.0
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)	1	▇ 0%	Apache-2.0
prior · pull toward the middle		▇ 0%
weighted average		= 45.8