whatmodel.sh / DeepSeek LLM 67B Chat

DeepSeek LLM 67B Chat

+ contribute

DeepSeek LLM 67B Chat is an open-weights language model from DeepSeek, released November 2023. Its published size is 67B parameters and a 4k-token context window. Licence: deepseek.

Catalogue identity notes

Comparison model from the Epoch AI internal-evaluation CSVs (Epoch release date 2023-11-29). Chat checkpoint deepseek-llm-67b-chat (November 2023). 67B parameters per the model name/card (no safetensors metadata).

40.8±1.3#518 of 719 LLMsfrom 11 benchmark results across 5 sources

$ how it compares

The LLMs ranked either side of DeepSeek LLM 67B Chat, out of 828 with a score, then more from DeepSeek.

#	model	score	where	RAM	released
608	DeepSeek-Coder 33B Base DeepSeek	41±1	local	~22GB	2023-10
609	InternLM2 Chat 20B Shanghai AI Laboratory (InternLM)	41±1	local	~14GB	2024-01
610	DeepSeek LLM 67B Chat DeepSeek	41±1	local	~31GB	2023-11
611	Llama 2 70B (base) Meta	41±1	local	~32GB	2023-07
612	Mixtral 8x7B v0.1 (base) Mistral AI	41±1	local	~29GB	2023-12
	more from DeepSeek
21	DeepSeek V4.1 Flash DeepSeek	78±1	api $0.26/M +local	~537GB	2026-09
31	DeepSeek V4 Pro (0813) DeepSeek	76±1	api $1.63/M +local	~1010GB	2026-08
35	DeepSeek V3.2 Speciale DeepSeek	76±3	comparison only

see all LLMs · find the ones that fit your machine

$ what it's good at

coding▁▁▁▃▆██▇▆▄▃▁ #252 of 458 · 8.6% of its score · 4 tests

Beats Qwen3-VL 32B Instruct, Solar Open 100B (102B-A12B), GLM-4.7-Flash here, though they rank higher overall.

test	result	vs the others in its run	difficulty	counts	why
CodeReviewQA: solution identification (hard), invariant accuracy	47.0% ✎	▃··▁▂▂▃▂█▃▄▂ beats 23/47	medium · avg 54%	▇ 3.4%	sharp · at its level · noisy
CodeReviewQA: change localisation (hard), invariant accuracy	37.8% ✎	█▂▁▁▂▂▂▃▅▅▅▇ beats 31/47	medium · avg 36%	▇ 2.6%	sharp · noisy
CodeReviewQA: automated code refinement, exact match rate	39.7% ✎	▁▁▁▁▁▁▂▂▃█▅▄ beats 39/47	hard · avg 33%	▇ 2.5%	typical
CodeReviewQA: change type recognition, invariant accuracy	79.1% ✎	▁▁▁······▁▃█ beats 45/47	easy · avg 67%	▇ 0.1%	blunt · at its level · noisy

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #322 of 393 · 72% of its score · 3 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	17.8% ✎	▁▂▂▃▆▆█▆▄▄▁▁ beats 47/122	medium · avg 39%	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 42%	typical
Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024)	1092.0 ✎	▂▁▂▁▃▇█▆▇▄▂▂ beats 42/95	avg 1201	▇▇▇▇▇▇ 15%	blunt
LMArena (text)	1184.9 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 63/379	avg 1268	▇▇▇▇▇▇ 15%	blunt

math▁▂▃▅▆▅▄█▄▁▁▁ #366 of 398 * · 11% of its score · 2 tests

Beats Yi 34B Chat, Command R (03-2024) here, though they rank higher overall. Trails Gemma 3 1B, OLMo 2 1B Instruct (0425), Stable Code 3B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
MATH level 5	6.4% ✎	▂▃▅▆▄█▆▃▅▂▃▄ beats 4/92	hard · avg 32%	▇▇▇▇ 10%	sharp · too hard for it
OTIS mock AIME 2024–2025	0.8% ✎	▃▃▄▃▃▃▇▇█▃▂▅ beats 7/193	very hard · avg 8.5%	▇ 0.8%	sharp · too hard for it · noisy

other tests2 not in a task · 7.2% of its score

test	result	vs the others in its run	difficulty	counts	why
GPQA Diamond	24.6% ✎	▃▁▁▁▂▄▆▄▅▇█▅ beats 2/208	hard · avg 40%	▇▇▇ 7.0%	sharp · too hard for it
Chess Puzzles (Epoch)	0.0% ✎	█▁▁▁▁▂▁▄▂▂▁▁ beats 15/135	very hard · avg 5.9%	▇ 0.2%	sharp · too hard for it · noisy

$ where to run it

llama.cpp · recommended

version

Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
Windows
irm https://llama.app/install.ps1 | iex
other ways to install ↗
Run DeepSeek LLM 67B Chat · needs ~40.4 GB · from huggingface.co/TheBloke
chat in the terminal
llama cli -hf TheBloke/deepseek-llm-67b-chat-GGUF:Q4_K_M
Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf TheBloke/deepseek-llm-67b-chat-GGUF:Q4_K_M

LM Studio

version

Install LM Studio · Open the app once before using the lms command.
download it (macOS / Windows / Linux) ↗
Run DeepSeek LLM 67B Chat · needs ~40.4 GB · from huggingface.co/TheBloke
open in LM Studio ↗
Or download it in the terminal
lms get https://huggingface.co/TheBloke/deepseek-llm-67b-chat-GGUF@Q4_K_M -y
Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

Install Ollama
download it (macOS / Windows) ↗
Linux
curl -fsSL https://ollama.com/install.sh | sh
Run DeepSeek LLM 67B Chat · needs ~40.7 GB · from ollama.com
chat in the terminal
ollama run deepseek-llm:67b-chat

Weights and documentation: huggingface.co/deepseek-ai ↗

$ how the score was worked out · AlpacaEval 1.0 and 2.0 leader… 42%, Epoch AI — internal evaluatio… 18%, Polyrating: A Cost-Effective… 15% and 2 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
AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)	1	▇▇▇▇▇▇▇▇▇▇▇▇▇ 42%	Apache-2.0
Epoch AI — internal evaluations	4	▇▇▇▇▇▇ 18%	CC BY 4.0
Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)	1	▇▇▇▇▇ 15%	CC-BY-4.0
LMArena leaderboard dataset	1	▇▇▇▇▇ 15%	CC BY 4.0
CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)	4	▇▇▇ 9%	CC-BY-4.0
prior · pull toward the middle		▇ 1%
weighted average		= 40.8