The Curious Engineer

All blogs / AI / ML

Prime Intellect

primeintellect.ai/blog

Train, deploy, and continuously improve your own models on an integrated compute, training, inference, and sandbox stack.

AI / MLDistributed systems / cloud

’24’25’26
Posts
58 from 23 Apr 2024 to 23 Oct 2026
Posting
weekly (39 in the last year)
Kind
company blog

2026 38

  1. AnnouncementsAPR 23RD, 2024State-of-the-Art in Globally-Distributed Training
  2. AnnouncementsJUL 1ST, 2024Introducing Prime Intellect Compute: The compute exchange
  3. AnnouncementsJUL 11TH, 2024OpenDiLoCo: An Open-Source Framework for Globally Distributed Low-Communication Training
  4. AnnouncementsOCT 11TH, 2024INTELLECT–1: Launching the First Globally-Distributed Training of a 10B Parameter Model
  5. AnnouncementsNOV 29TH, 2024INTELLECT-1 Release The First Globally Trained 10B Parameter Model
  6. AnnouncementsJAN 6TH, 2025METAGENE-1: Metagenomic Foundation Model
  7. AnnouncementsJAN 21ST, 2025Distributed Training in the Inference-Time-Compute Paradigm
  8. AnnouncementsJAN 28TH, 2025TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference
  9. AnnouncementsFEB 6TH, 2025SYNTHETIC-1: Scaling Distributed Synthetic Data Generation for Verified Reasoning
  10. AnnouncementsFEB 20TH, 2025SYNTHETIC-1 Release: Two Million Collaboratively Generated Reasoning Traces from Deepseek-R1
  11. AnnouncementsFEB 28TH, 2025$15M to Build The Open Superintelligence Stack
  12. AnnouncementsAPR 15TH, 2025INTELLECT-2: The First Globally Distributed Reinforcement Learning Training of a 32B Parameter Model
  13. AnnouncementsAPR 28TH, 2025Planetary-Scale Inference: Building a Distributed Inference Engine for the Public Internet
  14. AnnouncementsMAY 20TH, 2025PCCL - Prime Collective Communications Library
  15. AnnouncementsMAY 11TH, 2025INTELLECT-2 Release: The First 32B Parameter Model Trained Through Globally Distributed Reinforcement Learning
  16. AnnouncementsJUN 23RD, 2025SYNTHETIC-2
  17. AnnouncementsJUL 10TH, 2025SYNTHETIC-2 Release: Four Million Collaboratively Generated Reasoning Traces
  18. AnnouncementsAUG 27TH, 2025Environments Hub: A Community Hub To Scale RL To Open AGI
  19. AnnouncementsOCT 27TH, 2025Scaling Our Open-Source Environments Program
  20. AnnouncementsNOV 26TH, 2025INTELLECT-3: A 100B+ MoE trained with large-scale RL
  21. ResearchJAN 1ST, 2026Recursive Language Models: the paradigm of 2026
  22. AnnouncementsFEB 10TH, 2026Introducing Lab: The Full-Stack Platform for Training your Own Models
  23. PartnershipsMAR 16TH, 2026Leveraging NVIDIA to Build the Open Superintelligence Stack
  24. PartnershipsAPR 16TH, 2026Launching FrontierSWE on the Environments Hub
  25. PartnershipsMAR 30TH, 2026Partnering with Browserbase to Train Browser and Computer Use Agents
  26. Case StudyAPR 20TH, 2026How Zapier Turned AutomationBench Into a Continuous Agent Improvement Loop
  27. Case StudyMAY 7TH, 2026How Ramp Used RL to Beat Frontier Models at Spreadsheet Search
  28. AnnouncementsMAY 7TH, 2026Releasing Lab: the training platform for self-improving agents
  29. ResearchMAY 12TH, 2026renderers: Token-Level Templating for Agentic RL
  30. ResearchMAY 14TH, 2026Autonomous AI research for nanogpt speedrun
  31. ResearchMAY 20TH, 2026Systematic Reward Hacking and Prime Sprints
  32. ResearchMAY 18TH, 2026General Agent: A Self-Evolving, Synthetic Agent Environment
  33. AnnouncementsMAY 30TH, 2026Releasing Hosted Evaluations: Making benchmarking effortless
  34. PartnershipsJUN 4TH, 2026Prime Intellect Joins the NVIDIA Nemotron Coalition to Advance Open Frontier Models
  35. PartnershipsJUN 4TH, 2026Post-Training Nemotron 3 on Lab
  36. ResearchJUN 10TH, 2026True Agents Model the World
  37. ResearchJUN 22ND, 2026RL at 1T Scale: prime-rl Performance Deep Dive
  38. ResearchJUL 6TH, 2026prime-rl gets an Algorithms layer

2025 15

  1. AnnouncementsJUL 8TH, 2026$130M Series A to Build the Open Superintelligence Stack
  2. ResearchJUL 12TH, 2026verifiers v1: Decomposing Tasksets and Harnesses for Agentic RL & Evaluations
  3. ResearchJUL 22ND, 2026Scaling Agentic RL: 365,000+ Environments for SWE, Terminal, and Search
  4. ResearchAUG 5TH, 2026Prime Agent: A self-improving RLM agent
  5. ResearchAUG 7TH, 2026Multi-Agent Systems in PRIME-RL
  6. Case StudyAUG 13TH, 2026How Mixedbread Used Prime-RL to Beat Frontier Models in Deep Search
  7. ResearchAUG 13TH, 2026Prime Flash MoE - Faster MoE Kernels optimized for Blackwell
  8. ResearchAUG 15TH, 2026Measuring Autonomous AI Research
  9. ResearchAUG 25TH, 2026Uncovering a universal offline sandbox escape
  10. ResearchSEP 3RD, 2026GLM-5.2 RL weight transfer in 4 seconds using NIXL and ModelExpress
  11. Case StudySEP 17TH, 2026How Goodfire Uses Prime Intellect to Catch AI Models Cheating at Scale
  12. AnnouncementsSEP 23RD, 2026Prime Sandboxes: MicroVMs for Agentic RL Training at Scale
  13. Case StudyOCT 1ST, 2026How Extropic Uses Prime Intellect to Train a Thermodynamic ML Research Agent
  14. AnnouncementsOCT 2ND, 2026Prime Inference: Fast, Reliable Serving for Frontier Open Models
  15. ResearchOCT 6TH, 2026On the Nature of the Swarm

2024 5

  1. ResearchOCT 9TH, 2026Rewriting Prime Agent in Rust
  2. AnnouncementsOCT 13TH, 2026Moondream is joining Prime Intellect
  3. Case StudyOCT 15TH, 2026Post-Training Nemotron 3.5 Super VL for Long-Horizon Enterprise Work
  4. ResearchOCT 19TH, 2026prime-rl 0.10.0: faster training, our new RL loss IPO, and one unified eval stack
  5. ResearchOCT 23RD, 2026Prime Mega MoE: Fast Fused MoE Kernels for Training