whatmodel.sh / Command R (03-2024)

Command R (03-2024)

+ contribute

Command R (03-2024) is an open-weights language model from Cohere, released March 2024. Its published size is 35B parameters and a 128k-token context window. Licence: cc-by-nc-4.0.

Catalogue identity notes

Arena entry 'command-r' (text). LMArena's leaderboard table (arena-leaderboard space @6f2eeeb) labels 'command-r' as 'Command R (04-2024)' (Cohere, CC-BY-NC-4.0), linking https://txt.cohere.com/command-r. Maker repo CohereLabs/c4ai-command-r-v01 @760ddb6: licence tag cc-by-nc-4.0. 34,980,831,232 parameters from the Hugging Face safetensors metadata (dense, so active = total). Released 2024-03: month the maker's Hugging Face repo was created (createdAt 2024-03-11). Arena key 'command-r' (first Command R, open weights c4ai-command-r-v01); LMArena's table labels it 'Command R (04-2024)'. Distinct from command-r-08-2024. Cohere deprecations: effective September 15, 2025 'command-r-03-2024 (and the alias command-r)' deprecated.

41.6±0.8#506 of 719 LLMsfrom 15 benchmark results across 12 sources

$ how it compares

The LLMs ranked either side of Command R (03-2024), out of 828 with a score, then more from Cohere.

#	model	score	where	RAM	released
595	Granite 34B Code Base (8K) IBM	42±1	local	~24GB	2024-04
596	speechless-starcoder2-7b uukuguy	42±1	local	~6.1GB	2024-03
597	Command R (03-2024) Cohere	42±1	local	~24GB	2024-03
598	Ministral 3B (2410) Mistral	42±1	comparison only		2024-10
599	OLMo 2 7B Instruct (1124) Allen Institute for AI	42±1	local	~5.8GB	2024-12
	more from Cohere
76	North Small Translate 1.0 Cohere	71±6	api price unknown		2026-08
175	Command A+ (05-2026) Cohere	64±1	api price unknown +local	~71GB	2026-05
277	Command A (03-2025) Cohere	57±1	api $4.38/M +local	~37GB	2025-03

see all LLMs · find the ones that fit your machine

$ what it's good at

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #292 of 393 · 19% of its score · 3 tests

Beats Ministral 3 14B Reasoning 2512, Gemma 4 E4B, OpenThinker3 7B here, though they rank higher overall. Trails Flan-T5 XXL, LFM2 700M, EXAONE 3.5 2.4B Instruct here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena-Hard (v0.1, GPT-4-1106 judge)	17.0% ✎	▂▁▁▅█▅▄▃▃▅▃▁ beats 10/60	medium · avg 52%	▇▇▇▇ 9.2%	sharp
Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024)	1175.0 ✎	▂▁▂▁▃▇█▆▇▄▂▂ beats 75/95	avg 1201	▇▇ 5.1%	blunt
LMArena (text)	1226.8 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 82/379	avg 1268	▇▇ 5.0%	blunt

math▁▃▃▅▆▅▄█▄▁▁▁ #385 of 398 * · 12% of its score · 1 test

Trails Qwen2 0.5B Instruct, Granite 3.1 1B-A400M Instruct, Gemma 2B (base) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
GSM8K (ZeroEval, zero-shot)	53.0% ✎	▁▄·▁▄█▅▅▅███ beats 2/43	very easy · avg 89%	▇▇▇▇▇ 12%	sharp · at its level

other tests11 not in a task · 69% of its score

test	result	vs the others in its run	difficulty	counts	why
MMLU-Redux (ZeroEval, 2,778 questions)	61.1% ✎	▁·▁▁▄▄▄▅█▄▂▁ beats 3/53	easy · avg 77%	▇▇▇▇▇▇▇▇▇▇▇ 26%	at its level
LiveBench 2024-07-26	26.8% ✎	▁▁▁▂▂▃▃█▆▃▅▃ beats 20/61	medium · avg 41%	▇▇▇▇▇▇▇ 17%	at its level
MMLU-Pro	36.6% ✎	▃▂▅█▃·█▃▅··▂ beats 5/23	medium · avg 57%	▇▇▇ 7.9%	at its level
WildBench v2 WB-Score (task macro)	29.5331 WB-Score ✎	▁·▁▁·▃█▄▇▆▅▆ beats 12/52	avg 45.199 WB-Score	▇▇ 5.7%	sharp
ZebraLogic (ZeroEval, 1,000 puzzles)	9.9% ✎	▁▁▃▆██▁▁·▁▁▁ beats 9/57	hard · avg 18%	▇ 3.0%	sharp · too hard for it
Polish EQ-Bench v2 (SpeakLeash)	56.4% ✎	▁··▁▁▂▂▆▅▆█▆ beats 33/70	medium · avg 64%	▇ 3.0%	at its level
Open Persian LLM Leaderboard v2 average (21 tasks)	55.4% ✎	▂▄▂▃█▄▃▆▃▅▂▂ beats 44/81	medium · avg 58%	▇ 2.2%	blunt · at its level
RULER 64k	84.9% ✎	▁▁·▁·▂▃▄█▂·▁ beats 15/22	easy · avg 83%	▇ 1.5%	at its level

3 more tests

test	result	vs the others in its run	difficulty	counts	why
Open PT-LLM Leaderboard average (9 Portuguese tasks)	68.3% ✎	▃▁▁▁▁▁▁▁▂█▅▁ beats 151/276	medium · avg 73%	▇ 1.1%	at its level · noisy
EuroEval Life in the UK (English knowledge MC), test split, 5-shot, accuracy	84.3% ✎	▁▁▁▁▁▁▁▂▄█▄▃ beats 93/146	easy · avg 83%	▇ 0.8%	at its level · noisy
RULER 128k	76.0% ✎	▄··▂▄▂▆██▄▄▂ beats 14/21	easy · avg 68%	▇ 0.5%	at its level · 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 Command R (03-2024) · needs ~21.5 GB · from huggingface.co/lmstudio-community
chat in the terminal
llama cli -hf lmstudio-community/c4ai-command-r-v01-GGUF:Q4_K_M
Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf lmstudio-community/c4ai-command-r-v01-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 Command R (03-2024) · needs ~21.5 GB · from huggingface.co/lmstudio-community
open in LM Studio ↗
Or download it in the terminal
lms get https://huggingface.co/lmstudio-community/c4ai-command-r-v01-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 Command R (03-2024) · needs ~21.5 GB · from ollama.com
chat in the terminal
ollama run command-r:35b-v0.1-q4_K_M

MLX (Mac)

Install MLX (Mac) · Apple silicon Macs only, text only.
with uv
uv tool install mlx-lm
or
pip install mlx-lm
Run Command R (03-2024) · needs ~22.7 GB, Mac only · from huggingface.co/mlx-community
chat in the terminal
mlx_lm.chat --model mlx-community/c4ai-command-r-v01-4bit
Or start an OpenAI-compatible server at http://localhost:8080/v1
mlx_lm.server --model mlx-community/c4ai-command-r-v01-4bit

Weights and documentation: huggingface.co/CohereLabs ↗

$ how the score was worked out · WildEval ZeroEval published r… 41%, LiveBench model judgments (HF… 17%, Arena-Hard-Auto v0.1 leaderbo… 9% and 9 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
WildEval ZeroEval published results	3	▇▇▇▇▇▇▇▇▇▇▇▇ 41%	Apache-2.0
LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasets	1	▇▇▇▇▇ 17%	Apache-2.0 (stated in the LiveBench datasheet)
Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)	1	▇▇▇ 9%	Apache-2.0
MMLU-Pro author-released evaluation runs	1	▇▇ 8%	Apache-2.0
WildBench v2 leaderboard data (Ai2)	1	▇▇ 6%	Apache-2.0
Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)	1	▇▇ 5%	CC-BY-4.0
LMArena leaderboard dataset	1	▇▇ 5%	CC BY 4.0
SpeakLeash Polish EQ-Bench leaderboard	1	▇ 3%	Apache-2.0
Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)	1	▇ 2%	Apache-2.0
NVIDIA RULER repository README results table	2	▇ 2%	Apache-2.0
Open Portuguese LLM Leaderboard (archived)	1	▇ 1%	Apache-2.0
EuroEval leaderboards (results collection), English generative leaderboard	1	▇ 1%	MIT
prior · pull toward the middle		▇ 0%
weighted average		= 41.6

$ improve this page

Update model information Add a place to run it Add a test result Propose a benchmark Correct a benchmark score Edit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated 2026-10-11 02:20 UTC · sources and licences · json