Vai al contenuto
AI.info

tools

Core42

Core42 provides sovereign AI cloud infrastructure, model APIs, GPU compute, and services for enterprises, governments, and researchers.

Core42

In inglese

Core42 offers sovereign AI infrastructure for training, fine-tuning, and deploying AI workloads. Its products include Sovereign AI Cloud, Compass APIs and Playground, Sovereign Public Cloud, Signature Private Cloud, and delivery services.

Organizations use it for regulated workloads, government services, enterprise AI applications, research, and high-performance computing. Usage is priced by GPU hours, tokens, API calls, pages, or other consumption metrics; custom models and deployments may require contacting sales.

Features

  • Run AI workloads on NVIDIA, AMD, Cerebras, and Qualcomm accelerators
  • Use Compass APIs for text, embeddings, audio, vision, and real-time chat
  • Experiment with models in the Compass Playground
  • Train, fine-tune, and deploy models on sovereign cloud infrastructure
  • Use Kubernetes or Slurm for workload orchestration
  • Access on-demand GPU instances and large-scale clusters
  • Apply data residency, encryption, governance, and access controls

Use cases

  • Train and deploy models for regulated industries
  • Build government services with sovereign data controls
  • Run enterprise RAG, agent, and generative AI applications
  • Prototype AI applications through Compass Playground
  • Accelerate frontier AI and scientific research workloads

Pros

    Cons

      Latest updates

      • K2 Horizon 375B (5.5.2)

        Onboarded in the UAE region for advanced coding, complex problem-solving, and long-context agentic applications.

      • Pay-as-you-go Commercial Updates (5.5)

        Enhanced the Pay-as-you-go commercial experience to support subscription and monthly spend management for Azure Marketplace customers.

      • Cost by API Key in Usage Report (5.5)

        Usage Report now provides cost breakdowns by API key to track spending, allocate costs, and monitor usage.

      • GPT 5.5, GPT-5.6 Sol and Terra Deployment Update (5.4.1)

        Available in Sweden Central through TPM and Pay-as-you-go billing; supports Chat Completions API, Responses API, and Compass Playground.

      • GPT-5.6 Sol (5.4)

        Now available on Compass in the Sweden Central region, with support for the Chat Completions API and the Responses API.

      Capabilities

      • Choice of models — “unified access to frontier and open-weight models across multiple silicon types” source
      • Self-hosted — “The cloud enables flexible, secure AI deployment on premises and in the field, while cybersecurity ensures The Intelligence Grid remains guarded against threats.” source
      • API — “Use Compass APIs” source
      • Runs models for you — “A full-stack platform delivering accelerator choice, global scale, and peak performance across inference and training workloads.” source
      • Search over your data — “File Search Tool Call” source

      Get it

      Pricing

      Prices checked
      2026-09-25

      K2 Horizon 375B

      • $0.15 (input, per 1M tokens)
      • $0.50 (output, per 1M tokens)

        GPT-5.6 Sol

        • $5.50 (input, per 1M tokens)
        • $33.00 (output, per 1M tokens)

          GPT-5.6 Luna

          • $1.10 (input, per 1M tokens)
          • $6.60 (output, per 1M tokens)

            GPT-5.6 Terra

            • $2.75 (input, per 1M tokens)
            • $16.50 (output, per 1M tokens)

              GPT-5.4

              • $3.00 (input, per 1M tokens)
              • $18.15 (output, per 1M tokens)

                GPT-5.2

                • $1.75 (input, per 1M tokens)
                • $14.00 (output, per 1M tokens)

                  GPT-5.1

                  • $1.25 (input, per 1M tokens)
                  • $10.00 (output, per 1M tokens)

                    GPT-5

                    • $1.25 (input, per 1M tokens)
                    • $10.00 (output, per 1M tokens)

                      GPT-5 mini

                      • $0.275 (input, per 1M tokens)
                      • $2.20 (output, per 1M tokens)

                        GPT-4.1

                        • $2.00 (input, per 1M tokens)
                        • $8.00 (output, per 1M tokens)
                        • $30.3 (training, per 1M tokens)
                          Official website