oh, i need to try this
Prime Agent: a general purpose coding agent harness by Prime Intellect that uses RLM so that agents programmatically process their own near-infinite context. Based on Pi
github.com/PrimeIntelle...
Prime Agent: a general purpose coding agent harness by Prime Intellect that uses RLM so that agents programmatically process their own near-infinite context. Based on Pi
github.com/PrimeIntelle...
GitHub - PrimeIntellect-ai/prime-agent: Prime Agent is a self-improving RLM agent that can be used for coding workflows, or even long-running autonomous tasks.
Prime Agent is a self-improving RLM agent that can be used for coding workflows, or even long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
August 5, 2026 at 8:24 PM
oh, i need to try this
Prime Agent: a general purpose coding agent harness by Prime Intellect that uses RLM so that agents programmatically process their own near-infinite context. Based on Pi
github.com/PrimeIntelle...
Prime Agent: a general purpose coding agent harness by Prime Intellect that uses RLM so that agents programmatically process their own near-infinite context. Based on Pi
github.com/PrimeIntelle...
prime-agent is rust now lol github.com/PrimeIntelle...
Port Prime Agent to Rust (#2524) · PrimeIntellect-ai/prime-agent@39bc99a
## The Rust port of Prime Agent
A complete Rust rewrite of Prime Agent: the agent loop, the TUI, the daemon, the kernel orchestration, the ACP transport, the catalog system, and the full parity su...
github.com
October 1, 2026 at 12:12 PM
prime-agent is rust now lol github.com/PrimeIntelle...
Really excited about this one :)
Prime Agent + Opus 5 gets 95.5% on ARC-AGI-3 (179/183).
A big goal when I was building Prime Agent was making long-horizon agents more capable yet token efficient.
github.com/PrimeIntelle...
Prime Agent + Opus 5 gets 95.5% on ARC-AGI-3 (179/183).
A big goal when I was building Prime Agent was making long-horizon agents more capable yet token efficient.
github.com/PrimeIntelle...
GitHub - PrimeIntellect-ai/prime-agent: A self-improving RLM agent for coding workflows and long-running autonomous tasks.
A self-improving RLM agent for coding workflows and long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
August 10, 2026 at 6:23 PM
Really excited about this one :)
Prime Agent + Opus 5 gets 95.5% on ARC-AGI-3 (179/183).
A big goal when I was building Prime Agent was making long-horizon agents more capable yet token efficient.
github.com/PrimeIntelle...
Prime Agent + Opus 5 gets 95.5% on ARC-AGI-3 (179/183).
A big goal when I was building Prime Agent was making long-horizon agents more capable yet token efficient.
github.com/PrimeIntelle...
(optimus) prime iroh! github.com/PrimeIntelle...
GitHub - PrimeIntellect-ai/prime-iroh: Asynchronous P2P communication backend for decentralized pipeline parallelism
Asynchronous P2P communication backend for decentralized pipeline parallelism - PrimeIntellect-ai/prime-iroh
github.com
April 30, 2025 at 7:55 PM
(optimus) prime iroh! github.com/PrimeIntelle...
github.com/PrimeIntelle... its pretty buggy in some ways but this has been my favorite pi fork since they found a way to have the agent happily use a python REPL to self manage the way it manipulates data
GitHub - PrimeIntellect-ai/prime-agent: A self-improving RLM agent for coding workflows and long-running autonomous tasks.
A self-improving RLM agent for coding workflows and long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
August 29, 2026 at 2:41 PM
github.com/PrimeIntelle... its pretty buggy in some ways but this has been my favorite pi fork since they found a way to have the agent happily use a python REPL to self manage the way it manipulates data
Excited to release METAGENE-1, a 7B parameter metagenomic foundation model, built to aid in pathogen detection & pandemic monitoring. Pretrained on 1.5 trillion base pairs of DNA/RNA sequenced from wastewater.
A collab w/ USC, PrimeIntellect, & the Nucleic Acid Observatory.
metagene.ai
A collab w/ USC, PrimeIntellect, & the Nucleic Acid Observatory.
metagene.ai
Metagenomic Foundation Model
Metagenomic Foundation Model for Pandemic Monitoring
metagene.ai
January 7, 2025 at 8:58 PM
Excited to release METAGENE-1, a 7B parameter metagenomic foundation model, built to aid in pathogen detection & pandemic monitoring. Pretrained on 1.5 trillion base pairs of DNA/RNA sequenced from wastewater.
A collab w/ USC, PrimeIntellect, & the Nucleic Acid Observatory.
metagene.ai
A collab w/ USC, PrimeIntellect, & the Nucleic Acid Observatory.
metagene.ai
Excited to see that PrimeIntellect/INTELLECT-1-Instruct is the first non-Ai2 model to train on parts of the Tulu 3 datasets/recipe. Took about 1 week 🚀
https://buff.ly/3Zjmako
https://buff.ly/3Zjmako
December 2, 2024 at 4:13 PM
Excited to see that PrimeIntellect/INTELLECT-1-Instruct is the first non-Ai2 model to train on parts of the Tulu 3 datasets/recipe. Took about 1 week 🚀
https://buff.ly/3Zjmako
https://buff.ly/3Zjmako
i hereby am requesting that @ai2.bsky.social make a 1T base model and have PrimeIntellect posttrain it
November 27, 2025 at 7:41 PM
i hereby am requesting that @ai2.bsky.social make a 1T base model and have PrimeIntellect posttrain it
PI doesn't have a presence on here iirc so I'll do them the honors: they just released INTELLECT-3, a custom and fully open-source post-training recipe and associated weights for the recipe applied to GLM-4.5-Air. beats SOTA for the size!
huggingface.co/PrimeIntelle...
huggingface.co/PrimeIntelle...
PrimeIntellect/INTELLECT-3 · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
November 27, 2025 at 5:16 AM
PI doesn't have a presence on here iirc so I'll do them the honors: they just released INTELLECT-3, a custom and fully open-source post-training recipe and associated weights for the recipe applied to GLM-4.5-Air. beats SOTA for the size!
huggingface.co/PrimeIntelle...
huggingface.co/PrimeIntelle...
📦 PrimeIntellect-ai / prime-agent
⭐ 21,440 (+51)
🗒 Rust
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
⭐ 21,440 (+51)
🗒 Rust
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
GitHub - PrimeIntellect-ai/prime-agent: A self-improving RLM agent for coding workflows and long-running autonomous tasks.
A self-improving RLM agent for coding workflows and long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
October 1, 2026 at 4:17 PM
📦 PrimeIntellect-ai / prime-agent
⭐ 21,440 (+51)
🗒 Rust
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
⭐ 21,440 (+51)
🗒 Rust
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
there's a few labs that aren't frontier labs, but are doing such gosh darn interesting stuff that you absolutely need to be watching them
my list:
* Ai2 — open source AI, datasets, etc
* PrimeIntellect — federated/distributed model training
* Kyutai — TTS & realtime audio
* ...
who else?
my list:
* Ai2 — open source AI, datasets, etc
* PrimeIntellect — federated/distributed model training
* Kyutai — TTS & realtime audio
* ...
who else?
July 9, 2025 at 3:29 PM
there's a few labs that aren't frontier labs, but are doing such gosh darn interesting stuff that you absolutely need to be watching them
my list:
* Ai2 — open source AI, datasets, etc
* PrimeIntellect — federated/distributed model training
* Kyutai — TTS & realtime audio
* ...
who else?
my list:
* Ai2 — open source AI, datasets, etc
* PrimeIntellect — federated/distributed model training
* Kyutai — TTS & realtime audio
* ...
who else?
💎 Hidden Gem！ 💎 (1000+ new stars)
📦 PrimeIntellect-ai / prime-agent
⭐ 5,120 (+2,271)
🗒 TypeScript
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
📦 PrimeIntellect-ai / prime-agent
⭐ 5,120 (+2,271)
🗒 TypeScript
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
GitHub - PrimeIntellect-ai/prime-agent: A self-improving RLM agent for coding workflows and long-running autonomous tasks.
A self-improving RLM agent for coding workflows and long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
August 8, 2026 at 3:16 PM
💎 Hidden Gem！ 💎 (1000+ new stars)
📦 PrimeIntellect-ai / prime-agent
⭐ 5,120 (+2,271)
🗒 TypeScript
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
📦 PrimeIntellect-ai / prime-agent
⭐ 5,120 (+2,271)
🗒 TypeScript
A self-improving RLM agent for coding workflows and long-running autonomous tasks.
PrimeIntellect have released their tech report on INTELLECT-1: t.co/8hnoTILaL3
The first open-source world-wide training of a 10B model. The underlying ML distributed algo is DiLoCo (arxiv.org/abs/2311.08105) but they also built tons of engineering on top of it to make it scalable.
The first open-source world-wide training of a 10B model. The underlying ML distributed algo is DiLoCo (arxiv.org/abs/2311.08105) but they also built tons of engineering on top of it to make it scalable.
December 1, 2024 at 10:10 AM
PrimeIntellect have released their tech report on INTELLECT-1: t.co/8hnoTILaL3
The first open-source world-wide training of a 10B model. The underlying ML distributed algo is DiLoCo (arxiv.org/abs/2311.08105) but they also built tons of engineering on top of it to make it scalable.
The first open-source world-wide training of a 10B model. The underlying ML distributed algo is DiLoCo (arxiv.org/abs/2311.08105) but they also built tons of engineering on top of it to make it scalable.
alright, i love Prime Agent, but it's been steadily becoming less stable
i finally sat down and fixed it this morning. Rewrote a fairly large chunk of it so comms don't get bottlenecked, which eliminates all the global failures that used to happen
it's pretty nice now
github.com/PrimeIntelle...
i finally sat down and fixed it this morning. Rewrote a fairly large chunk of it so comms don't get bottlenecked, which eliminates all the global failures that used to happen
it's pretty nice now
github.com/PrimeIntelle...
[General] Is v0.8.1 usable at all? · PrimeIntellect-ai prime-agent · Discussion #1805
Area Development Topic It seems like the stability of the harness has progressively gotten worse. It's to the point where the harness won't stay alive more than 5 minutes. People are creating featu...
github.com
August 30, 2026 at 5:17 PM
alright, i love Prime Agent, but it's been steadily becoming less stable
i finally sat down and fixed it this morning. Rewrote a fairly large chunk of it so comms don't get bottlenecked, which eliminates all the global failures that used to happen
it's pretty nice now
github.com/PrimeIntelle...
i finally sat down and fixed it this morning. Rewrote a fairly large chunk of it so comms don't get bottlenecked, which eliminates all the global failures that used to happen
it's pretty nice now
github.com/PrimeIntelle...
imo LoRAs might be a better answer than open source RL environments like PrimeIntellect
at the end, you have a small weights file that you can run alongside lots of other LoRAs, or distribute just weights
idk just feels like LoRAs are far more economically useful
at the end, you have a small weights file that you can run alongside lots of other LoRAs, or distribute just weights
idk just feels like LoRAs are far more economically useful
afaict SF/SV has hyped RL environments to the extent that everyone is building them in excess
kinda like MCP all over
it’s been going on for a couple months, and now we’re expecting several big model releases in the next month or two
it’ll be interesting to see what happens
kinda like MCP all over
it’s been going on for a couple months, and now we’re expecting several big model releases in the next month or two
it’ll be interesting to see what happens
i think October may have some interesting AI releases
October 1, 2025 at 11:14 PM
imo LoRAs might be a better answer than open source RL environments like PrimeIntellect
at the end, you have a small weights file that you can run alongside lots of other LoRAs, or distribute just weights
idk just feels like LoRAs are far more economically useful
at the end, you have a small weights file that you can run alongside lots of other LoRAs, or distribute just weights
idk just feels like LoRAs are far more economically useful
On GitHub, PrimeIntellect-ai presents experiments with autonomous agents optimizing NN training, reflecting advancements in speedrunning. Over three waves, agents Claude and Codex exceeded benchmarks. https://github.com/PrimeIntellect-ai/experiments-autonomous-speedrunning/tree/main
GitHub - PrimeIntellect-ai/experiments-autonomous-speedrunning: autonomous nanogpt optimizer speedrun · GitHub
On GitHub, PrimeIntellect-ai presents experiments with autonomous agents optimizing NN training, reflecting advancements in speedrunning. Over three waves, agents Claude and Codex exceeded benchmarks.
github.com
June 18, 2026 at 5:20 AM
On GitHub, PrimeIntellect-ai presents experiments with autonomous agents optimizing NN training, reflecting advancements in speedrunning. Over three waves, agents Claude and Codex exceeded benchmarks. https://github.com/PrimeIntellect-ai/experiments-autonomous-speedrunning/tree/main
is distributed training the future?
PrimeIntellect is assembling the world’s first fully distributed training run of a 32B model. Anyone can participate with whatever hardware they have available. Fully permissionless.
www.primeintellect.ai/blog/intelle...
PrimeIntellect is assembling the world’s first fully distributed training run of a 32B model. Anyone can participate with whatever hardware they have available. Fully permissionless.
www.primeintellect.ai/blog/intelle...
INTELLECT-2: The First Globally Distributed Reinforcement Learning Training of a 32B Parameter Model
Today we are launching INTELLECT-2: the first 32B parameter globally decentralized Reinforcement Learning training run where anyone can permissionlessly contribute their heterogeneous compute resource...
www.primeintellect.ai
April 15, 2025 at 11:08 PM
is distributed training the future?
PrimeIntellect is assembling the world’s first fully distributed training run of a 32B model. Anyone can participate with whatever hardware they have available. Fully permissionless.
www.primeintellect.ai/blog/intelle...
PrimeIntellect is assembling the world’s first fully distributed training run of a 32B model. Anyone can participate with whatever hardware they have available. Fully permissionless.
www.primeintellect.ai/blog/intelle...
can i convince yall to look at the repl that github.com/PrimeIntelle... uses, it really does seem like it lets the agent work in a much more token efficient way from my testing compared to omp
GitHub - PrimeIntellect-ai/prime-agent: A self-improving RLM agent for coding workflows and long-running autonomous tasks.
A self-improving RLM agent for coding workflows and long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
August 31, 2026 at 6:17 AM
can i convince yall to look at the repl that github.com/PrimeIntelle... uses, it really does seem like it lets the agent work in a much more token efficient way from my testing compared to omp
github.com/PrimeIntelle... Got some buzz right now
GitHub - PrimeIntellect-ai/prime-agent: A self-improving RLM agent for coding workflows and long-running autonomous tasks.
A self-improving RLM agent for coding workflows and long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
August 12, 2026 at 6:40 PM
github.com/PrimeIntelle... Got some buzz right now
RT @np_hard: As a follow-up on my previous multi-agent work for the RL Residency @PrimeIntellect, I've been exploring cooperative agents with communication (aka swarms). I wrote-up my findings and thoughts in this post: https://nphard.io/2026/10/07/eleusinian-swarms.html
October 7, 2026 at 8:11 PM
RT @np_hard: As a follow-up on my previous multi-agent work for the RL Residency @PrimeIntellect, I've been exploring cooperative agents with communication (aka swarms). I wrote-up my findings and thoughts in this post: https://nphard.io/2026/10/07/eleusinian-swarms.html
How do they get the info? 😂 Eulether, HuggingFace, PrimeIntellect, Pi, a lot more
May 9, 2026 at 4:09 PM
How do they get the info? 😂 Eulether, HuggingFace, PrimeIntellect, Pi, a lot more
Really excited about this one :)
Prime Agent + Opus 5 gets 95.5% on ARC-AGI-3 (179/183).
A big goal when I was building Prime Agent was making long-horizon agents more capable yet token efficient.
github.com/PrimeIntelle...
Prime Agent + Opus 5 gets 95.5% on ARC-AGI-3 (179/183).
A big goal when I was building Prime Agent was making long-horizon agents more capable yet token efficient.
github.com/PrimeIntelle...
GitHub - PrimeIntellect-ai/prime-agent: A self-improving RLM agent for coding workflows and long-running autonomous tasks.
A self-improving RLM agent for coding workflows and long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
August 10, 2026 at 6:23 PM
Really excited about this one :)
Prime Agent + Opus 5 gets 95.5% on ARC-AGI-3 (179/183).
A big goal when I was building Prime Agent was making long-horizon agents more capable yet token efficient.
github.com/PrimeIntelle...
Prime Agent + Opus 5 gets 95.5% on ARC-AGI-3 (179/183).
A big goal when I was building Prime Agent was making long-horizon agents more capable yet token efficient.
github.com/PrimeIntelle...
1/ Prime Intellect DiLoCo leads the pack. Just trained a 10B parameter model using decentralized compute! Their web UI at app.primeintellect.ai makes it easy for anyone to join in. Truly democratizing AI compute. Check the OSS project: github.com/PrimeIntellect-ai/OpenDiloco (2/9)
PrimeIntellect Compute | Advanced AI Chips at Best Prices
Discover PrimeIntellect’s unbeatable deals on state-of-the-art AI compute solutions
app.primeintellect.ai
January 2, 2025 at 12:33 AM
1/ Prime Intellect DiLoCo leads the pack. Just trained a 10B parameter model using decentralized compute! Their web UI at app.primeintellect.ai makes it easy for anyone to join in. Truly democratizing AI compute. Check the OSS project: github.com/PrimeIntellect-ai/OpenDiloco (2/9)
I've been really liking prime-agent with luna max lately. The amount of work you can squeeze out of the $20 subscription is kinda nuts. github.com/PrimeIntelle...
GitHub - PrimeIntellect-ai/prime-agent: A self-improving RLM agent for coding workflows and long-running autonomous tasks.
A self-improving RLM agent for coding workflows and long-running autonomous tasks. - PrimeIntellect-ai/prime-agent
github.com
August 26, 2026 at 2:51 PM
I've been really liking prime-agent with luna max lately. The amount of work you can squeeze out of the $20 subscription is kinda nuts. github.com/PrimeIntelle...