---
type: "raw_signal"
id: "introducing-olmo-core-3-open-scalable-training-infrastructure-97cb3a"
slug: "introducing-olmo-core-3-open-scalable-training-infrastructure-97cb3a"
title: "Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs"
description: "Ai2 has released Olmo-core 3, an open training framework redesigned for large mixture-of-experts (MoE) models."
retrieval_nugget: "Ai2 has released Olmo-core 3, an open training framework redesigned for large mixture-of-experts (MoE) models."
observed_at: "2026-10-01"
record_date: "2026-10-01"
date_kind: "observed_at"
why_it_matters: "Ai2 has released Olmo-core 3, an open training framework redesigned for large mixture-of-experts (MoE) models. In a preliminary test on eight NVIDIA B300 GPUs, a 47-billion-parameter MoE processed 52,000 tokens per second per GPU, versus 19,400 with Ai2’s earlier implementation, or about 2.7 times the throughput. A separate benchmark increased the expert pool from 8 to 128 while selecting four exp"
novelty: "notable"
verification_level: "source-linked"
signal_type: "release"
source_platform: "allenai.org"
topics: []
entities: []
related_patterns: []
source_url: "https://allenai.org/blog/olmocore3"
source_urls: ["https://allenai.org/blog/olmocore3"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:introducing-olmo-core-3-open-scalable-training-infrastructure-97cb3a"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/signals/introducing-olmo-core-3-open-scalable-training-infrastructure-97cb3a/","markdown":"https://newruntime.com/signals/introducing-olmo-core-3-open-scalable-training-infrastructure-97cb3a.md","json":"https://newruntime.com/signals/introducing-olmo-core-3-open-scalable-training-infrastructure-97cb3a.json"}
---

# Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs

## Retrieval answer

Ai2 has released Olmo-core 3, an open training framework redesigned for large mixture-of-experts (MoE) models.

Ai2 has released Olmo-core 3, an open training framework redesigned for large mixture-of-experts (MoE) models. In a preliminary test on eight NVIDIA B300 GPUs, a 47-billion-parameter MoE processed 52,000 tokens per second per GPU, versus 19,400 with Ai2’s earlier implementation, or about 2.7 times the throughput. A separate benchmark increased the expert pool from 8 to 128 while selecting four experts per token: total capacity rose from 4.6 billion to 47 billion parameters while throughput fell by less than 5%. Ai2 also reports a 1.2-trillion-parameter configuration benchmarked across 512 GPUs. Those tests used random routing to measure system performance, not the quality of a trained model; a 2.38-trillion-parameter configuration was a short-capacity test, not a full training run. The stack is intended for Ai2’s next Olmo generation and is being released for outside use.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Raw Signals](https://newruntime.com/signals.json): Open every source-linked raw signal.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)