January 15, 2025
Farabi Center Enters Top 600 Supercomputers Globally
About 2 PFLOPS peak performance places the centre among the world’s stronger systems.
al-Farabi Kazakh National University
About 2 PFLOPS of peak performance, 430 servers, and 6 PB of data storage. Farabi Supercomputer Center at al-Farabi Kazakh National University provides HPC, cloud, and AI services for science, government, and business.
Peak Performance
CPU Compute Nodes
NVIDIA A10 Accelerators
Storage
InfiniBand Network
What we run
Batch science jobs, self-service VMs, and AI inference — without stitching three vendors together.
Parallel jobs for simulation, modelling, and large scientific workloads.
VPS, VDS, and VDI, with optional GPU passthrough when you need it.
Deploy common research and ops apps from a curated catalogue.
Managed databases and storage with backups and high availability.
OpenAI-compatible API, Kazakh LLM work, RAG, and embeddings.
Linux mirrors, public Git, package registries, and DNS helpers.
Services
From bare-metal nodes to managed AI inference.
400+ Intel Xeon CPU nodes, InfiniBand HDR, and 6 PB storage for large scientific jobs.
60× NVIDIA A10 GPUs for training, inference, and vision workloads.
Self-service VMs with standard OS images; GPU passthrough available.
PostgreSQL, Redis, MongoDB, Elasticsearch, MinIO, and RabbitMQ with HA and backups.
JupyterHub, GitLab, WordPress, Grafana, and 30+ other packaged apps.
Help with job scripts, cluster use, cloud layout, and AI deployment on Farabi.
Research on Farabi
Partners
Latest News
January 15, 2025
About 2 PFLOPS peak performance places the centre among the world’s stronger systems.
December 8, 2024
ISSAI KAZ-LLM-8B reached 52% on Kazakh benchmarks, ahead of GPT-4o on native Kazakh tasks.
November 20, 2024
Self-service VMs are available through the portal, with optional GPU passthrough.