[whatmodel.sh](../../README.md) / Command A (03-2025) # Command A (03-2025) Cohere · released 2025-03 · 111B params · cc-by-nc-4.0 **score 57.9** ±1.0 · #228 of 732 · from 22 benchmark results across 6 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at _{● marks Command A (03-2025) among every model on that task or test, worst to best.}

task            vs the others          rank  counts  tests
agent tasks     ▁▁▃▆●▄▅▅▄▄▂▁    #198 of 313     28%      6
coding          ▁▁▁▃▅▇●▇▅▄▃▁    #230 of 469     38%      5
chat & writing  ▁▁▁▃▆▄▆█●▇▇▃   #161 of 402*     10%      2
math            ▁▁▃▅█●▄█▄▂▁▁   #256 of 431*    8.5%      2
web search      ▁▁▁▅▄●▃▄·▁·▁     #46 of 84*    4.4%      1

agent tasks · #198 of 313 · 28% of its score

Beats [o1 (2024-12-17)](o1-2024-12-17.md), [o3-mini](o3-mini.md), [DeepSeek R1 0528](deepseek-r1-0528.md) here, though they rank higher overall. Trails [SmolLM3 3B](smollm3-3b.md), [Jamba 1.5 Mini](jamba-1-5-mini.md), [Mellum2 12B-A2.5B Instruct](mellum2-12b-a2-5b-instruct.md) here, though they rank lower overall.

test                                result  vs the others in its run   difficulty            counts  why
SWE-bench Verified (maker)           26.8%  ▃···▃·▃·●█·▃ beats 3/7     hard · avg 17%           12%  sharp · at its level
BFCL V4 (December 2025 archive)      46.5%  ▁▁▁▂██▅●▆▂▂▂ beats 52/77   hard · avg 31%          4.6%  at its level
Finance Agent Benchmark v1 (537 q…    4.6%  █·●▃·█▅▅▅▅▃▅ beats 3/18    very hard · avg 5.5%    4.4%  sharp
τ-bench Airline (maker)              45.3%  ▁▁▁·▅▄▁●▇▁▁▇ beats 14/28   hard · avg 22%          2.7%  at its level
BFCL v3                              64.4%  ▁▁▁▁▁▃▅█▅●▁▁ beats 95/114  medium · avg 51%        2.1%  blunt · at its level
τ-bench Retail (maker)               60.0%  ▁····▃·▂●█▂▆ beats 9/28    hard · avg 27%          1.9%  at its level · noisy

coding · #230 of 469 · 38% of its score

Beats [Qwen3.5 4B](qwen3-5-4b.md), [Qwen3-VL 32B Instruct](qwen3-vl-32b-instruct.md), [Solar Open 100B (102B-A12B)](solar-open-100b.md) here, though they rank higher overall. Trails [QwQ-32B-Preview](qwq-32b-preview.md), [Ministral 3 3B Reasoning 2512](ministral-3-3b-reasoning.md), [GPT-4 (0613)](gpt-4-0613.md) here, though they rank lower overall.

test                                result  vs the others in its run  difficulty            counts  why
Aider polyglot (225 exercises, tw…   14.7%  █▄··█·●····▄ beats 6/7    very hard · avg 5.5%     12%  sharp · noisy
SWE-bench Verified (maker)           26.8%  ▃···▃·▃·●█·▃ beats 3/7    hard · avg 17%           12%  sharp · at its level
LiveCodeBench v5 (Oct 2024–Feb 20…   26.9%  ▄▄·▄·····●▄█ beats 4/7    hard · avg 18%          6.7%  at its level
MBPP+ (maker)                        86.2%  ██·██···●███ beats 4/7    easy · avg 78%          3.7%  typical
BIRD-SQL (dev) (maker)               59.5%  ▄··▄·▄·█··▄● beats 7/7    medium · avg 47%        3.3%  at its level

chat & writing · #161 of 402 · 10% of its score

Beats [MiniMax-M2](minimax-m2.md), [MiniCPM5 2B](minicpm5-2b.md), [Granite 4.2 30B](granite-4-2-30b.md) here, though they rank higher overall. Trails [LFM2.5 1.2B Thinking](lfm2-5-1-2b-thinking.md), [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md), [LFM2.5 2.6B](lfm2-5-2-6b.md) here, though they rank lower overall.

test                                result  vs the others in its run    difficulty  counts  why
LMArena (text)                      1353.7  ▁▁▁▂▃▃▃▆●▇█▃ beats 199/385  avg 1260       10%  blunt
IFEval strict, mean of prompt and…   90.9%  █··········● beats 1/1                       —  not counted yet: no other model in the …

math · #256 of 431 · 8.5% of its score

Trails [DeepSeek R1 Distill Qwen 1.5B](deepseek-r1-distill-qwen-1-5b.md), [Gemma 7B (base)](gemma-7b-base.md), [CodeGemma 7B (v1.0)](codegemma-7b.md) here, though they rank lower overall.

test               result  vs the others in its run  difficulty           counts  why
AIME 2024 (maker)   23.3%  ▁▁▁▂▁●▁▅███▃ beats 13/85  very hard · avg 14%    8.5%  sharp · at its level · small run
MATH (maker)        80.0%  ▃▅▂▆▅▃▅▅█▃▃● beats 30/30  medium · avg 61%          —  not counted yet: no other model in the …

web search · #46 of 84 · 4.4% of its score

Beats [Gemini 3.1 Pro (Preview)](gemini-3-1-pro.md), [Qwen3.7 Max](qwen3-7-max.md), [GPT-5.2](gpt-5-2.md) here, though they rank higher overall. Trails [GPT-4o mini (2024-07-18)](gpt-4o-mini-2024-07-18.md), [Mistral Small 3.1 24B (2503)](mistral-small-3-1.md), [Llama 4 Scout (109B-A17B)](llama-4-scout.md) here, though they rank lower overall.

test                                result  vs the others in its run  difficulty            counts  why
Finance Agent Benchmark v1 (537 q…    4.6%  █·●▃·█▅▅▅▅▃▅ beats 3/18   very hard · avg 5.5%    4.4%  sharp

other tests (11)

test                                result  vs the others in its run   difficulty           counts  why
RepoQA (32K, threshold 0.8) (make…   92.6%  ▄▄▄··▄▄··▄·● beats 7/7     easy · avg 80%         4.8%  typical
BigCodeBench (Instruct) (maker)      45.4%  ▄··▄▄···▄▄●▄ beats 5/7     hard · avg 35%         4.5%  at its level
GPQA Diamond (simple-evals) (make…   50.8%  ▁·····▄●▂▂▃▃ beats 11/22   very hard · avg 35%    4.3%  sharp · at its level · small run
Open Persian LLM Leaderboard v2 a…   70.0%  ▂▄▂▃█▄▃▆▃●▂▂ beats 72/81   medium · avg 57%       4.1%  blunt · at its level
MMLU-Pro (maker)                     69.6%  ▁▁·▁▁▁▃▂●▅█▁ beats 86/188  medium · avg 54%       3.5%  at its level · small run
LongBench v2 (maker)                 43.4%  ▂·▁·▂▃▁▄●▃█▄ beats 13/33   very hard · avg 33%    3.2%  small run
RULER 128k (maker)                   90.0%  ▃··▅▄█▅█▇·●▃ beats 31/33   easy · avg 70%         1.7%  small run
MMLU (simple-evals, zero-shot CoT…   85.5%  ▁···▂▁▅●▂▂▃▁ beats 9/22    easy · avg 76%         1.5%  small run
Vectara hallucination leaderboard…   95.5%  ▁▁▂▃▄▄●▆▃▂▂▁ beats 74/140  very easy · avg 94%    0.4%  blunt · maxed out · noisy
Vectara hallucination leaderboard…   90.7%  ▁▂▁▂▅●▂▃▄▃▁▁ beats 56/98   very easy · avg 90%    0.0%  blunt
InFoBench (maker)                    94.9%  █··········● beats 1/1                               —  not counted yet: no other model in the …

## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's 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 is added up by who ran the tests. [Method](../method.md)

source                                    results  share of the score
Cohere (submitted / maker-reported)            15   74% ▇▇▇▇▇▇▇▇▇
LMArena leaderboard dataset                     1   10% ▇
Berkeley Function Calling Leaderboard r…        2    7% ▇
Finance Agent Benchmark                         1    4% ▇
Open Persian LLM Leaderboard v2                 1    4% ▇
Vectara hallucination leaderboard               2    0% ▇
prior: pull toward the middle                        0% ▇
weighted average                                   = 57.9

_{Licences: Cohere ([CC BY 4.0](../../licenses/report-2504.00698v2.txt)) · LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Berkeley Function Calling Leaderboard result archive ([Apache-2.0](../../licenses/bfcl-result-archive-v4-Apache-2.0.txt)) · Finance Agent Benchmark ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · Open Persian LLM Leaderboard v2 ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))}

Not counted (3)

source  test                            score  why
Cohere  IFEval strict, mean of prompt…  90.9%  too few models have it yet
Cohere  InFoBench                       94.9%  too few models have it yet
Cohere  MATH                            80.0%  too few models have it yet

## How to run it ### Through an API api (Microsoft Azure) · $4.38/M blended · $2.5 in / $10 out per 1M tokens · [azure.microsoft.com](https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/cohere) _{[price source](https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/cohere): USD per 1M tokens, Azure AI Foundry Models pay-as-you-go, 'Cohere Command A Global' (Azure model id cohere-command-a, version 1, GA with no retirement date on Azure's schedule): input $2.50, output $10.00, region US East. Cohere's own pricing page lists no Command A per-token rate. Currency: USD. Microsoft Azure official pricing page (price cells carry the amounts in the page's own data), observed 2026-09-25. · [context source](https://docs.cohere.com/docs/models): Cohere API model table: command-a-03-2025 context length 256k (stated rounded), maximum output 8k tokens. Observed 2026-09-25.} ### llama.cpp (recommended) 1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md) ```sh # macOS / Linux curl -LsSf https://llama.app/install.sh | sh # Windows irm https://llama.app/install.ps1 | iex ``` 2. Run Command A (03-2025) (Q2_K, needs ~33.8 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/c4ai-command-a-03-2025-GGUF/tree/main)) ```sh llama cli -hf unsloth/c4ai-command-a-03-2025-GGUF:Q2_K ``` 3. Or start a local server and open the address it prints (http://localhost:9931) ```sh llama serve -hf unsloth/c4ai-command-a-03-2025-GGUF:Q2_K ``` ### LM Studio 1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download) 2. Run Command A (03-2025) (4bit, needs ~62.5 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/c4ai-command-a-03-2025-4bit/tree/a4c0a1f6fd2a921bdb5f3a77f855ee9f44850fb5)) ```sh lms get https://huggingface.co/mlx-community/c4ai-command-a-03-2025-4bit -y ``` 3. Then run lms ls to see its name, and lms chat followed by that name Other versions: Q2_K (~33.8 GB) `lms get https://huggingface.co/unsloth/c4ai-command-a-03-2025-GGUF@Q2_K -y` · 8bit (~118 GB, Mac only) `lms get https://huggingface.co/mlx-community/c4ai-command-a-03-2025-8bit -y` ### Ollama 1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download) ```sh # Linux curl -fsSL https://ollama.com/install.sh | sh ``` 2. Run Command A (03-2025) (Q4_K_M, needs ~67.1 GB, from [ollama.com](https://ollama.com/library/command-a/tags)) ```sh ollama run command-a:111b ``` ### MLX (Mac) 1. Install MLX (Mac). Apple silicon Macs only, text only. ```sh # with uv uv tool install mlx-lm # or pip install mlx-lm ``` 2. Run Command A (03-2025) (4bit, needs ~62.5 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/c4ai-command-a-03-2025-4bit/tree/a4c0a1f6fd2a921bdb5f3a77f855ee9f44850fb5)) ```sh mlx_lm.chat --model mlx-community/c4ai-command-a-03-2025-4bit ``` 3. Or start an OpenAI-compatible server at http://localhost:8080/v1 ```sh mlx_lm.server --model mlx-community/c4ai-command-a-03-2025-4bit ``` Other versions: 8bit (~118 GB, Mac only) `mlx_lm.chat --model mlx-community/c4ai-command-a-03-2025-8bit` Model repository and documentation: [huggingface.co/CohereLabs](https://huggingface.co/CohereLabs/c4ai-command-a-03-2025) [interactive version](https://whatmodel.sh/model/command-a) · [JSON](https://whatmodel.sh/api?format=json&all=1)