Image de couverture de livre

Command A: An Enterprise-Ready Large Language Model

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Cohere, Team, :, Aakanksha, Ahmadian, Arash, Ahmed, Marwan, Alammar, Jay, Alizadeh, Milad, Alnumay, Yazeed, Althammer, Sophia, Arkhangorodsky, Arkady, Aryabumi, Viraat, Aumiller, Dennis, Avalos, Raphaël, Aviv, Zahara, Bae, Sammie, Baji, Saurabh, Barbet, Alexandre, Bartolo, Max, Bebensee, Björn, Beladia, Neeral, Beller-Morales, Walter, Bérard, Alexandre, Berneshawi, Andrew, Bialas, Anna, Blunsom, Phil, Bobkin, Matt, Bongale, Adi, Braun, Sam, Brunet, Maxime, Cahyawijaya, Samuel, Cairuz, David, Campos, Jon Ander, Cao, Cassie, Cao, Kris, Castagné, Roman, Cendrero, Julián, Currie, Leila Chan, Chandak, Yash, Chang, Diane, Chatziveroglou, Giannis, Chen, Hongyu, Cheng, Claire, Chevalier, Alexis, Chiu, Justin T., Cho, Eugene, Choi, Eugene, Choi, Eujeong, Chung, Tim, Cirik, Volkan, Cismaru, Ana, Clavier, Pierre, Conklin, Henry, Crawhall-Stein, Lucas, Crouse, Devon, Cruz-Salinas, Andres Felipe, Cyrus, Ben, D'souza, Daniel, Dalla-Torre, Hugo, Dang, John, Darling, William, Domingues, Omar Darwiche, Dash, Saurabh, Debugne, Antoine, Dehaze, Théo, Desai, Shaan, Devassy, Joan, Dholakia, Rishit, Duffy, Kyle, Edalati, Ali, Eldeib, Ace, Elkady, Abdullah, Elsharkawy, Sarah, Ergün, Irem, Ermis, Beyza, Fadaee, Marzieh, Fan, Boyu, Fayoux, Lucas, Flet-Berliac, Yannis, Frosst, Nick, Gallé, Matthias, Galuba, Wojciech, Garg, Utsav, Geist, Matthieu, Azar, Mohammad Gheshlaghi, Gilsenan-McMahon, Ellen, Goldfarb-Tarrant, Seraphina, Goldsack, Tomas, Gomez, Aidan, Gonzaga, Victor Machado, Govindarajan, Nithya, Govindassamy, Manoj, Grinsztajn, Nathan, Gritsch, Nikolas, Gu, Patrick, Guo, Shangmin, Haefeli, Kilian, Hajjar, Rod, Hawes, Tim, He, Jingyi, Hofstätter, Sebastian, Hong, Sungjin, Hooker, Sara, Hosking, Tom, Howe, Stephanie, Hu, Eric, Huang, Renjie, Jain, Hemant, Jain, Ritika, Jakobi, Nick, Jenkins, Madeline, Jordan, JJ, Joshi, Dhruti, Jung, Jason, Kalyanpur, Trushant, Kamalakara, Siddhartha Rao, Kedrzycki, Julia, Keskin, Gokce, Kim, Edward, Kim, Joon, Ko, Wei-Yin, Kocmi, Tom, Kozakov, Michael, Kryściński, Wojciech, Jain, Arnav Kumar, Teru, Komal Kumar, Land, Sander, Lasby, Michael, Lasche, Olivia, Lee, Justin, Lewis, Patrick, Li, Jeffrey, Li, Jonathan, Lin, Hangyu, Locatelli, Acyr, Luong, Kevin, Ma, Raymond, Mach, Lukáš, Machado, Marina, Magbitang, Joanne, Lopez, Brenda Malacara, Mann, Aryan, Marchisio, Kelly, Markham, Olivia, Matton, Alexandre, McKinney, Alex, McLoughlin, Dominic, Mokry, Jozef, Morisot, Adrien, Moulder, Autumn, Moynehan, Harry, Mozes, Maximilian, Muppalla, Vivek, Murakhovska, Lidiya, Nagarajan, Hemangani, Nandula, Alekhya, Nasir, Hisham, Nehra, Shauna, Netto-Rosen, Josh, Ohashi, Daniel, Owers-Bardsley, James, Ozuzu, Jason, Padilla, Dennis, Park, Gloria, Passaglia, Sam, Pekmez, Jeremy, Penstone, Laura, Piktus, Aleksandra, Ploeg, Case, Poulton, Andrew, Qi, Youran, Raghvendra, Shubha, Ramos, Miguel, Ranjan, Ekagra, Richemond, Pierre, Robert-Michon, Cécile, Rodriguez, Aurélien, Roy, Sudip, Ruder, Sebastian, Ruis, Laura, Rust, Louise, Sachan, Anubhav, Salamanca, Alejandro, Saravanakumar, Kailash Karthik, Satyakam, Isha, Sebag, Alice Schoenauer, Sen, Priyanka, Sepehri, Sholeh, Seshadri, Preethi, Shen, Ye, Sherborne, Tom, Shi, Sylvie Shang, Shivaprasad, Sanal, Shmyhlo, Vladyslav, Shrinivason, Anirudh, Shteinbuk, Inna, Shukayev, Amir, Simard, Mathieu, Snyder, Ella, Spataru, Ava, Spooner, Victoria, Starostina, Trisha, Strub, Florian, Su, Yixuan, Sun, Jimin, Talupuru, Dwarak, Tarassov, Eugene, Tommasone, Elena, Tracey, Jennifer, Trend, Billy, Tumer, Evren, Üstün, Ahmet, Venkitesh, Bharat, Venuto, David, Verga, Pat, Voisin, Maxime, Wang, Alex, Wang, Donglu, Wang, Shijian, Wen, Edmond, White, Naomi, Willman, Jesse, Winkels, Marysia, Xia, Chen, Xie, Jessica, Xu, Minjie, Yang, Bowen, Yi-Chern, Tan, Zhang, Ivan, Zhao, Zhenyu, Zhao, Zhoujie
Format:	Preprint
Publié:	2025
Sujets:	Computation and Language Artificial Intelligence Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

_version_	1866915239646199808
author	Cohere, Team : Aakanksha Ahmadian, Arash Ahmed, Marwan Alammar, Jay Alizadeh, Milad Alnumay, Yazeed Althammer, Sophia Arkhangorodsky, Arkady Aryabumi, Viraat Aumiller, Dennis Avalos, Raphaël Aviv, Zahara Bae, Sammie Baji, Saurabh Barbet, Alexandre Bartolo, Max Bebensee, Björn Beladia, Neeral Beller-Morales, Walter Bérard, Alexandre Berneshawi, Andrew Bialas, Anna Blunsom, Phil Bobkin, Matt Bongale, Adi Braun, Sam Brunet, Maxime Cahyawijaya, Samuel Cairuz, David Campos, Jon Ander Cao, Cassie Cao, Kris Castagné, Roman Cendrero, Julián Currie, Leila Chan Chandak, Yash Chang, Diane Chatziveroglou, Giannis Chen, Hongyu Cheng, Claire Chevalier, Alexis Chiu, Justin T. Cho, Eugene Choi, Eugene Choi, Eujeong Chung, Tim Cirik, Volkan Cismaru, Ana Clavier, Pierre Conklin, Henry Crawhall-Stein, Lucas Crouse, Devon Cruz-Salinas, Andres Felipe Cyrus, Ben D'souza, Daniel Dalla-Torre, Hugo Dang, John Darling, William Domingues, Omar Darwiche Dash, Saurabh Debugne, Antoine Dehaze, Théo Desai, Shaan Devassy, Joan Dholakia, Rishit Duffy, Kyle Edalati, Ali Eldeib, Ace Elkady, Abdullah Elsharkawy, Sarah Ergün, Irem Ermis, Beyza Fadaee, Marzieh Fan, Boyu Fayoux, Lucas Flet-Berliac, Yannis Frosst, Nick Gallé, Matthias Galuba, Wojciech Garg, Utsav Geist, Matthieu Azar, Mohammad Gheshlaghi Gilsenan-McMahon, Ellen Goldfarb-Tarrant, Seraphina Goldsack, Tomas Gomez, Aidan Gonzaga, Victor Machado Govindarajan, Nithya Govindassamy, Manoj Grinsztajn, Nathan Gritsch, Nikolas Gu, Patrick Guo, Shangmin Haefeli, Kilian Hajjar, Rod Hawes, Tim He, Jingyi Hofstätter, Sebastian Hong, Sungjin Hooker, Sara Hosking, Tom Howe, Stephanie Hu, Eric Huang, Renjie Jain, Hemant Jain, Ritika Jakobi, Nick Jenkins, Madeline Jordan, JJ Joshi, Dhruti Jung, Jason Kalyanpur, Trushant Kamalakara, Siddhartha Rao Kedrzycki, Julia Keskin, Gokce Kim, Edward Kim, Joon Ko, Wei-Yin Kocmi, Tom Kozakov, Michael Kryściński, Wojciech Jain, Arnav Kumar Teru, Komal Kumar Land, Sander Lasby, Michael Lasche, Olivia Lee, Justin Lewis, Patrick Li, Jeffrey Li, Jonathan Lin, Hangyu Locatelli, Acyr Luong, Kevin Ma, Raymond Mach, Lukáš Machado, Marina Magbitang, Joanne Lopez, Brenda Malacara Mann, Aryan Marchisio, Kelly Markham, Olivia Matton, Alexandre McKinney, Alex McLoughlin, Dominic Mokry, Jozef Morisot, Adrien Moulder, Autumn Moynehan, Harry Mozes, Maximilian Muppalla, Vivek Murakhovska, Lidiya Nagarajan, Hemangani Nandula, Alekhya Nasir, Hisham Nehra, Shauna Netto-Rosen, Josh Ohashi, Daniel Owers-Bardsley, James Ozuzu, Jason Padilla, Dennis Park, Gloria Passaglia, Sam Pekmez, Jeremy Penstone, Laura Piktus, Aleksandra Ploeg, Case Poulton, Andrew Qi, Youran Raghvendra, Shubha Ramos, Miguel Ranjan, Ekagra Richemond, Pierre Robert-Michon, Cécile Rodriguez, Aurélien Roy, Sudip Ruder, Sebastian Ruis, Laura Rust, Louise Sachan, Anubhav Salamanca, Alejandro Saravanakumar, Kailash Karthik Satyakam, Isha Sebag, Alice Schoenauer Sen, Priyanka Sepehri, Sholeh Seshadri, Preethi Shen, Ye Sherborne, Tom Shi, Sylvie Shang Shivaprasad, Sanal Shmyhlo, Vladyslav Shrinivason, Anirudh Shteinbuk, Inna Shukayev, Amir Simard, Mathieu Snyder, Ella Spataru, Ava Spooner, Victoria Starostina, Trisha Strub, Florian Su, Yixuan Sun, Jimin Talupuru, Dwarak Tarassov, Eugene Tommasone, Elena Tracey, Jennifer Trend, Billy Tumer, Evren Üstün, Ahmet Venkitesh, Bharat Venuto, David Verga, Pat Voisin, Maxime Wang, Alex Wang, Donglu Wang, Shijian Wen, Edmond White, Naomi Willman, Jesse Winkels, Marysia Xia, Chen Xie, Jessica Xu, Minjie Yang, Bowen Yi-Chern, Tan Zhang, Ivan Zhao, Zhenyu Zhao, Zhoujie
author_facet	Cohere, Team : Aakanksha Ahmadian, Arash Ahmed, Marwan Alammar, Jay Alizadeh, Milad Alnumay, Yazeed Althammer, Sophia Arkhangorodsky, Arkady Aryabumi, Viraat Aumiller, Dennis Avalos, Raphaël Aviv, Zahara Bae, Sammie Baji, Saurabh Barbet, Alexandre Bartolo, Max Bebensee, Björn Beladia, Neeral Beller-Morales, Walter Bérard, Alexandre Berneshawi, Andrew Bialas, Anna Blunsom, Phil Bobkin, Matt Bongale, Adi Braun, Sam Brunet, Maxime Cahyawijaya, Samuel Cairuz, David Campos, Jon Ander Cao, Cassie Cao, Kris Castagné, Roman Cendrero, Julián Currie, Leila Chan Chandak, Yash Chang, Diane Chatziveroglou, Giannis Chen, Hongyu Cheng, Claire Chevalier, Alexis Chiu, Justin T. Cho, Eugene Choi, Eugene Choi, Eujeong Chung, Tim Cirik, Volkan Cismaru, Ana Clavier, Pierre Conklin, Henry Crawhall-Stein, Lucas Crouse, Devon Cruz-Salinas, Andres Felipe Cyrus, Ben D'souza, Daniel Dalla-Torre, Hugo Dang, John Darling, William Domingues, Omar Darwiche Dash, Saurabh Debugne, Antoine Dehaze, Théo Desai, Shaan Devassy, Joan Dholakia, Rishit Duffy, Kyle Edalati, Ali Eldeib, Ace Elkady, Abdullah Elsharkawy, Sarah Ergün, Irem Ermis, Beyza Fadaee, Marzieh Fan, Boyu Fayoux, Lucas Flet-Berliac, Yannis Frosst, Nick Gallé, Matthias Galuba, Wojciech Garg, Utsav Geist, Matthieu Azar, Mohammad Gheshlaghi Gilsenan-McMahon, Ellen Goldfarb-Tarrant, Seraphina Goldsack, Tomas Gomez, Aidan Gonzaga, Victor Machado Govindarajan, Nithya Govindassamy, Manoj Grinsztajn, Nathan Gritsch, Nikolas Gu, Patrick Guo, Shangmin Haefeli, Kilian Hajjar, Rod Hawes, Tim He, Jingyi Hofstätter, Sebastian Hong, Sungjin Hooker, Sara Hosking, Tom Howe, Stephanie Hu, Eric Huang, Renjie Jain, Hemant Jain, Ritika Jakobi, Nick Jenkins, Madeline Jordan, JJ Joshi, Dhruti Jung, Jason Kalyanpur, Trushant Kamalakara, Siddhartha Rao Kedrzycki, Julia Keskin, Gokce Kim, Edward Kim, Joon Ko, Wei-Yin Kocmi, Tom Kozakov, Michael Kryściński, Wojciech Jain, Arnav Kumar Teru, Komal Kumar Land, Sander Lasby, Michael Lasche, Olivia Lee, Justin Lewis, Patrick Li, Jeffrey Li, Jonathan Lin, Hangyu Locatelli, Acyr Luong, Kevin Ma, Raymond Mach, Lukáš Machado, Marina Magbitang, Joanne Lopez, Brenda Malacara Mann, Aryan Marchisio, Kelly Markham, Olivia Matton, Alexandre McKinney, Alex McLoughlin, Dominic Mokry, Jozef Morisot, Adrien Moulder, Autumn Moynehan, Harry Mozes, Maximilian Muppalla, Vivek Murakhovska, Lidiya Nagarajan, Hemangani Nandula, Alekhya Nasir, Hisham Nehra, Shauna Netto-Rosen, Josh Ohashi, Daniel Owers-Bardsley, James Ozuzu, Jason Padilla, Dennis Park, Gloria Passaglia, Sam Pekmez, Jeremy Penstone, Laura Piktus, Aleksandra Ploeg, Case Poulton, Andrew Qi, Youran Raghvendra, Shubha Ramos, Miguel Ranjan, Ekagra Richemond, Pierre Robert-Michon, Cécile Rodriguez, Aurélien Roy, Sudip Ruder, Sebastian Ruis, Laura Rust, Louise Sachan, Anubhav Salamanca, Alejandro Saravanakumar, Kailash Karthik Satyakam, Isha Sebag, Alice Schoenauer Sen, Priyanka Sepehri, Sholeh Seshadri, Preethi Shen, Ye Sherborne, Tom Shi, Sylvie Shang Shivaprasad, Sanal Shmyhlo, Vladyslav Shrinivason, Anirudh Shteinbuk, Inna Shukayev, Amir Simard, Mathieu Snyder, Ella Spataru, Ava Spooner, Victoria Starostina, Trisha Strub, Florian Su, Yixuan Sun, Jimin Talupuru, Dwarak Tarassov, Eugene Tommasone, Elena Tracey, Jennifer Trend, Billy Tumer, Evren Üstün, Ahmet Venkitesh, Bharat Venuto, David Verga, Pat Voisin, Maxime Wang, Alex Wang, Donglu Wang, Shijian Wen, Edmond White, Naomi Willman, Jesse Winkels, Marysia Xia, Chen Xie, Jessica Xu, Minjie Yang, Bowen Yi-Chern, Tan Zhang, Ivan Zhao, Zhenyu Zhao, Zhoujie
contents	In this report we describe the development of Command A, a powerful large language model purpose-built to excel at real-world enterprise use cases. Command A is an agent-optimised and multilingual-capable model, with support for 23 languages of global business, and a novel hybrid architecture balancing efficiency with top of the range performance. It offers best-in-class Retrieval Augmented Generation (RAG) capabilities with grounding and tool use to automate sophisticated business processes. These abilities are achieved through a decentralised training approach, including self-refinement algorithms and model merging techniques. We also include results for Command R7B which shares capability and architectural similarities to Command A. Weights for both models have been released for research purposes. This technical report details our original training pipeline and presents an extensive evaluation of our models across a suite of enterprise-relevant tasks and public benchmarks, demonstrating excellent performance and efficiency.
format	Preprint
id	arxiv_https___arxiv_org_abs_2504_00698
institution	arXiv
publishDate	2025
record_format	arxiv
spellingShingle	Command A: An Enterprise-Ready Large Language Model Cohere, Team : Aakanksha Ahmadian, Arash Ahmed, Marwan Alammar, Jay Alizadeh, Milad Alnumay, Yazeed Althammer, Sophia Arkhangorodsky, Arkady Aryabumi, Viraat Aumiller, Dennis Avalos, Raphaël Aviv, Zahara Bae, Sammie Baji, Saurabh Barbet, Alexandre Bartolo, Max Bebensee, Björn Beladia, Neeral Beller-Morales, Walter Bérard, Alexandre Berneshawi, Andrew Bialas, Anna Blunsom, Phil Bobkin, Matt Bongale, Adi Braun, Sam Brunet, Maxime Cahyawijaya, Samuel Cairuz, David Campos, Jon Ander Cao, Cassie Cao, Kris Castagné, Roman Cendrero, Julián Currie, Leila Chan Chandak, Yash Chang, Diane Chatziveroglou, Giannis Chen, Hongyu Cheng, Claire Chevalier, Alexis Chiu, Justin T. Cho, Eugene Choi, Eugene Choi, Eujeong Chung, Tim Cirik, Volkan Cismaru, Ana Clavier, Pierre Conklin, Henry Crawhall-Stein, Lucas Crouse, Devon Cruz-Salinas, Andres Felipe Cyrus, Ben D'souza, Daniel Dalla-Torre, Hugo Dang, John Darling, William Domingues, Omar Darwiche Dash, Saurabh Debugne, Antoine Dehaze, Théo Desai, Shaan Devassy, Joan Dholakia, Rishit Duffy, Kyle Edalati, Ali Eldeib, Ace Elkady, Abdullah Elsharkawy, Sarah Ergün, Irem Ermis, Beyza Fadaee, Marzieh Fan, Boyu Fayoux, Lucas Flet-Berliac, Yannis Frosst, Nick Gallé, Matthias Galuba, Wojciech Garg, Utsav Geist, Matthieu Azar, Mohammad Gheshlaghi Gilsenan-McMahon, Ellen Goldfarb-Tarrant, Seraphina Goldsack, Tomas Gomez, Aidan Gonzaga, Victor Machado Govindarajan, Nithya Govindassamy, Manoj Grinsztajn, Nathan Gritsch, Nikolas Gu, Patrick Guo, Shangmin Haefeli, Kilian Hajjar, Rod Hawes, Tim He, Jingyi Hofstätter, Sebastian Hong, Sungjin Hooker, Sara Hosking, Tom Howe, Stephanie Hu, Eric Huang, Renjie Jain, Hemant Jain, Ritika Jakobi, Nick Jenkins, Madeline Jordan, JJ Joshi, Dhruti Jung, Jason Kalyanpur, Trushant Kamalakara, Siddhartha Rao Kedrzycki, Julia Keskin, Gokce Kim, Edward Kim, Joon Ko, Wei-Yin Kocmi, Tom Kozakov, Michael Kryściński, Wojciech Jain, Arnav Kumar Teru, Komal Kumar Land, Sander Lasby, Michael Lasche, Olivia Lee, Justin Lewis, Patrick Li, Jeffrey Li, Jonathan Lin, Hangyu Locatelli, Acyr Luong, Kevin Ma, Raymond Mach, Lukáš Machado, Marina Magbitang, Joanne Lopez, Brenda Malacara Mann, Aryan Marchisio, Kelly Markham, Olivia Matton, Alexandre McKinney, Alex McLoughlin, Dominic Mokry, Jozef Morisot, Adrien Moulder, Autumn Moynehan, Harry Mozes, Maximilian Muppalla, Vivek Murakhovska, Lidiya Nagarajan, Hemangani Nandula, Alekhya Nasir, Hisham Nehra, Shauna Netto-Rosen, Josh Ohashi, Daniel Owers-Bardsley, James Ozuzu, Jason Padilla, Dennis Park, Gloria Passaglia, Sam Pekmez, Jeremy Penstone, Laura Piktus, Aleksandra Ploeg, Case Poulton, Andrew Qi, Youran Raghvendra, Shubha Ramos, Miguel Ranjan, Ekagra Richemond, Pierre Robert-Michon, Cécile Rodriguez, Aurélien Roy, Sudip Ruder, Sebastian Ruis, Laura Rust, Louise Sachan, Anubhav Salamanca, Alejandro Saravanakumar, Kailash Karthik Satyakam, Isha Sebag, Alice Schoenauer Sen, Priyanka Sepehri, Sholeh Seshadri, Preethi Shen, Ye Sherborne, Tom Shi, Sylvie Shang Shivaprasad, Sanal Shmyhlo, Vladyslav Shrinivason, Anirudh Shteinbuk, Inna Shukayev, Amir Simard, Mathieu Snyder, Ella Spataru, Ava Spooner, Victoria Starostina, Trisha Strub, Florian Su, Yixuan Sun, Jimin Talupuru, Dwarak Tarassov, Eugene Tommasone, Elena Tracey, Jennifer Trend, Billy Tumer, Evren Üstün, Ahmet Venkitesh, Bharat Venuto, David Verga, Pat Voisin, Maxime Wang, Alex Wang, Donglu Wang, Shijian Wen, Edmond White, Naomi Willman, Jesse Winkels, Marysia Xia, Chen Xie, Jessica Xu, Minjie Yang, Bowen Yi-Chern, Tan Zhang, Ivan Zhao, Zhenyu Zhao, Zhoujie Computation and Language Artificial Intelligence Machine Learning In this report we describe the development of Command A, a powerful large language model purpose-built to excel at real-world enterprise use cases. Command A is an agent-optimised and multilingual-capable model, with support for 23 languages of global business, and a novel hybrid architecture balancing efficiency with top of the range performance. It offers best-in-class Retrieval Augmented Generation (RAG) capabilities with grounding and tool use to automate sophisticated business processes. These abilities are achieved through a decentralised training approach, including self-refinement algorithms and model merging techniques. We also include results for Command R7B which shares capability and architectural similarities to Command A. Weights for both models have been released for research purposes. This technical report details our original training pipeline and presents an extensive evaluation of our models across a suite of enterprise-relevant tasks and public benchmarks, demonstrating excellent performance and efficiency.
title	Command A: An Enterprise-Ready Large Language Model
topic	Computation and Language Artificial Intelligence Machine Learning
url	https://arxiv.org/abs/2504.00698