Command R+
Command R+ is Cohere’s flagship RAG-optimised model, designed for enterprise retrieval workloads with strong tool use and multilingual support.
Released
April 4, 2024
Type
llm
License
open-weight
Context
128,000 tokens
Pricing
Input
$2.50 / 1M tokens
Output
$10.00 / 1M tokens
Capabilities
Architecture
Parameters: 104B
Links
Command R+ in the news
MarkTechPost · Jul 23, 2026
Best Open Speech Recognition (ASR) Models in 2026: WER, Languages, Latency, and License Compared
The landscape of open Automatic Speech Recognition (ASR) models has evolved beyond a single dominant player, with several models now competing on performance metrics and licensing. In March 2026, Cohere released Transcribe, a 2B parameter model under the Apache 2.0 license, which initially led the Hugging Face Open ASR Leaderboard with an average Word Error Rate (WER) of 5.42%. This was quickly followed by IBM's Granite Speech 4.1 2B, which achieved a WER of 5.33%. More recently, ARK-ASR-3B and MOSS-Transcribe-preview-2B have posted even lower WER scores, narrowing the gap at the top of the leaderboard to less than one WER point. This intense competition means that rank alone is no longer the primary factor for users selecting an ASR model. Instead, critical considerations now include the model's license, the breadth of languages it supports, its streaming capabilities for real-time processing, and the cost per audio hour. These factors are becoming increasingly important for developers and businesses integrating ASR technology into their applications. However, a closer examination of the Open ASR Leaderboard reveals nuances in how the reported average WER is calculated. The average WER cited is not a fixed quantity, and models have been scored using different test sets. For instance, Cohere's 5.42% WER is an average across eight English test sets, including TED-LIUM. In contrast, ARK-ASR-3B's 5.04% is an average across seven sets, excluding TED-LIUM. The MOSS-Transcribe-preview-2B card explicitly states that TED-LIUM is not part of its current leaderboard run. Since TED-LIUM is considered one of the easier test sets, its exclusion can artificially inflate the reported average WER. When Cohere's published per-dataset scores are recomputed using the same seven test sets that ARK reports, Cohere's average WER increases to 5.84% from the headline 5.42%. Similarly, IBM's Granite Speech 4.1 2B moves from 5.33% to 5.65% when evaluated on the same basis. This suggests that, on a like-for-like comparison, ARK's lead over other models might be larger than initially indicated by the headline leaderboard figures. The focus is shifting towards transparent and consistent evaluation methodologies to accurately compare these rapidly advancing open ASR models.
Hugging Face · Jul 16, 2026
NVIDIA Nemotron 3 Embed Ranks #1 Overall on RTEB, Advancing Agentic Retrieval
NVIDIA's Nemotron 3 Embed model achieved the top overall score on the Retrieval-Augmented Generation (RAG) Evaluation Benchmark (RTEB) this week, marking a significant advancement in agentic retrieval capabilities. The model outperformed previous leaders, including models from Cohere and Google, across the benchmark's diverse set of tasks designed to assess retrieval-augmented generation performance. The RTEB, developed by researchers at Peking University and the National University of Singapore, evaluates models on their ability to retrieve relevant information and integrate it into generated text. Nemotron 3 Embed's success demonstrates its enhanced capacity for understanding context and accurately sourcing information, crucial for building more capable AI agents. This benchmark is critical for the development of AI systems that can reliably access and utilize external knowledge bases. Nemotron 3 Embed's superior performance on RTEB suggests a leap forward in the development of AI agents that can perform complex tasks requiring both information retrieval and sophisticated language generation. The benchmark's comprehensive evaluation covers various aspects of RAG, including retrieval accuracy, relevance, and the seamless integration of retrieved information into coherent and contextually appropriate responses. NVIDIA's achievement highlights the growing importance of specialized embedding models in powering advanced AI applications. This ranking underscores the competitive landscape of AI model development, with Nemotron 3 Embed now setting a new standard for retrieval-augmented generation. The benchmark's results are expected to influence future research and development in the field, pushing for more robust and reliable AI systems. The ongoing evolution of benchmarks like RTEB is essential for tracking progress and identifying leading technologies in the rapidly advancing domain of artificial intelligence.
CoinTelegraph · Jun 18, 2026
HIVE secures $220M AI infrastructure contract with Bell and Cohere
HIVE Blockchain Technologies Ltd. secured a significant AI infrastructure contract valued at $220 million with Bell Canada and Cohere on May 28, 2024. This agreement is projected to contribute approximately $70 million in annual recurring revenue for HIVE as it expands its artificial intelligence-focused operations. The contract involves HIVE providing cloud computing infrastructure to support the AI development and deployment needs of both Bell and Cohere. Bell Canada, a major telecommunications company, is leveraging advanced AI for its services, while Cohere, an AI company, focuses on developing large language models. The partnership underscores the growing demand for specialized AI infrastructure and HIVE's strategic positioning to meet this demand. The company's investment in high-performance computing is expected to accelerate its growth in the AI sector.
CoinDesk · Jun 18, 2026
Hive shares jumps 10% on $220m Canada sovereign AI infrastructure deal
Hive Digital Technologies Ltd. experienced a 10% surge in its share price on March 13, 2024, following the announcement of a significant $220 million sovereign artificial intelligence (AI) infrastructure deal with the Canadian government. This multi-year contract positions Hive as a key provider of GPU cloud computing resources for Canada's national AI initiatives. The agreement involves Bell Canada and Cohere, a prominent AI company, as partners in delivering these services. This strategic move marks a substantial pivot for Hive, transitioning its business model away from its historical focus on Bitcoin mining towards high-performance computing essential for AI development and deployment. The deal underscores the growing demand for dedicated AI infrastructure and Hive's ambition to become a leading player in this rapidly expanding market. The company's stock saw its value increase by approximately $0.75 per share, reaching $8.25 following the news.