Cohere vs Mistral AI

Introducing North 2: AI without compromises

Cohere has announced North 2, its latest platform update with 15+ new features, positioned as enterprise AI without compromises.

North 2: Enterprise AI without compromises

Cohere's biggest upgrade yet — unifying enterprise-grade security, performance, and flexibility in one platform.

Aleph Alpha's Kolibri Lags Behind Open-Weight Leaders

Aleph Alpha has released Kolibri, a new open-source German-English AI model under the Apache-2.0 license, designed for mission-critical applications in public administration and industry. Kolibri features a Mixture-of-Experts architecture with 78 billion total parameters and about 3.46 billion active per token. Trained on 20 trillion tokens, with 23% in German and a German-optimized tokenizer, it supports context lengths up to one million tokens. The model excels in agentic workflows, RAG, and native tool calling, while emphasizing data sovereignty and compliance, including a blocklist of 4.5 million URLs and GDPR/EU AI Act adherence. Independent benchmarks, however, place it behind at least 20 other open-weight models. Aleph Alpha is merging with Cohere to form a transatlantic sovereign AI company, subject to regulatory approval. In a broader economic debate, Reid Hoffman argues AI infrastructure expansion is the 'only reason' the US avoids recession, though economists dispute this.

• Aleph Alpha released Kolibri under Apache-2.0, with 78B total parameters and ~3.46B active per token.
• Kolibri trained on 20T tokens, with 23% German data and a German-optimized tokenizer, supporting 1M context.
• Independent benchmarks rank Kolibri behind at least 20 other open-weight models.

Mistral Launches Timestral Time-Series Models, Competing with NXAI

Mistral AI has launched Timestral, its first family of time-series forecasting models, directly challenging Austria's NXAI and its TiRex model. Timestral comprises four variants with 5, 28, 123, and 209 million parameters, designed to run efficiently on standard server CPUs without GPUs, processing up to 32,768 historical data points and forecasting 1,024 steps ahead. Mistral claims the largest model ranks first on three common time-series leaderboards (GIFT-Eval, TIME, fev-bench), outperforming Google's TimesFM-3, Amazon's Chronos-2, and Datadog's Toto, as well as NXAI's TiRex-2. The models target industrial sectors like automotive, energy, and supply chain. Mistral had previously acquired Emmi AI, an NXAI spin-off, reportedly valuing it at up to €330 million. These results are self-reported and not yet independently verified.

• Mistral AI launched Timestral, a family of four time-series forecasting models with 5, 28, 123, and 209 million parameters.
• Timestral models run on standard server CPUs without GPUs, with inference times from 10 to 140 milliseconds.
• Mistral claims its largest Timestral model ranks first on GIFT-Eval, TIME, and fev-bench leaderboards among pre-trained models.

NVIDIA Expands European Investments, Stockholm AI Lab

NVIDIA has revealed a broad array of European activities ahead of GTC Berlin, including planned investments in nine startups across the UK and France, collaborations in Sweden, and the launch of a Stockholm AI Lab. The company is participating in funding rounds for startups such as Boltz, Foresight, Hiverge, Prima Mente, Riverlane, Doctolib, Entalpic, Genesis AI, and Pigment. NVIDIA also announced a partnership with the Norrsken Foundation for the new AI hub, and is collaborating with Sweden AI Factory to open infrastructure to startups. These investments are part of a wider strategy to deepen its presence in Europe's AI ecosystem, which has seen about €7.72 billion in funding rounds involving NVIDIA this year. NVIDIA's CEO Jensen Huang underscored Europe's industrial potential, while the company's moves also raise questions about technological dependence on US suppliers.

• NVIDIA announced planned investments in five UK startups: Boltz, Foresight, Hiverge, Prima Mente, Riverlane.
• NVIDIA announced planned minority investments in four French startups: Doctolib, Entalpic, Genesis AI, Pigment.
• NVIDIA and Norrsken Foundation will establish the Stockholm AI Lab, opening in 2027.

Mistral CEO Warns Europe: Frontier AI Dependency Risk

Mistral AI CEO Arthur Mensch argues that Europe must develop its own frontier AI models to avoid economic and strategic dependence on foreign providers. In a LinkedIn essay, he warns that relying on open models from the US and China locks Europe into dependency, citing a June incident where a US lab cut access to its models and the high costs of American cloud services. He highlights Mistral's plans to increase compute capacity sixfold and its recent €3 billion funding round at a €21 billion valuation as proof of Europe's ability to compete. However, critics, including EU Commission President Ursula von der Leyen and Ecosia's founder, advocate for an applications-first approach. Ecosia has dropped Mistral for cheaper Chinese open models, underscoring market challenges. The debate emphasizes the strategic importance of AI sovereignty in Europe.

• Mistral AI CEO Arthur Mensch published a LinkedIn essay urging Europe to build its own frontier AI models, warning against reliance on foreign open models.
• Mensch cites a June incident where a US lab cut access to its models and the high costs of American cloud services as evidence of dependency.
• Mistral plans to increase its compute capacity sixfold and recently raised €3 billion at a €21 billion valuation.