The Open-Source AI Stack
RSS
All models

Models · olmo

OLMo 2 7B Instruct

Open AI2 · 2024-11-26 · Apache-2.0

The only major open-weights release that meets the strictest reading of OSAID: weights, training data (Dolma), training code (OLMo trainer), and training logs (WandB). Reference for what "fully open" can mean in 2024.

Cost

— / Mtok input
— / Mtok output

Together AI · as of 2026-05-19

via Artificial Analysis ↗

Speed

— tok/sec output

Together AI · as of 2026-05-19

via Artificial Analysis ↗

Architecture

tokens in Embedding vocab not disclosed · olmo tokenizer × N layers Multi-head Attention RoPE context 4,096 tokens Dense MLP SwiGLU activation (standard) 7.3B active params Output projection tokens out
Schema-generated from data/models.yaml. Every label is auditable against the model's sources.

Specs

Architecture
dense
Total params
7B
Active params
7.3B
Context window
not verified
Attention
mha
Position encoding
rope
Pretraining tokens
4.0T
Training hardware
H100
Post-training
sft, dpo
OSI-approved
yes
Data released
yes
Training code
released

Benchmarks

Each score carries the date it was published; we never infer or interpolate missing scores.

General reasoning

MMLU 61.3 as of 2024-12-02 source ↗

Recommended use cases

  • research reproducibility
  • OSAID-compliant deployment
  • fine-tuning base

Available quantizations

GGUF llama.cpp's container; the common local format, k-quants from Q2 to Q8. runs on llama.cpp, Ollama
AWQ Activation-aware 4-bit weight quantization for GPU serving. runs on vLLM, SGLang
GPTQ Post-training 4-bit weight quantization for GPU serving. runs on vLLM, SGLang, Transformers
MLX Apple MLX 4/8-bit layout for Apple silicon. runs on Apple MLX

Verified via the Hugging Face model tree ↗. Community quantizations change over time; the families shown are those with published weights at audit time.

Notable innovations

  • · Full training stack open (data + code + logs)
  • · OSAID-compliant

Known limitations

  • · 5T-token pretrain is well below the 15-36T used by 2025-class open-weights releases; benchmark scores reflect this. source ↗

Lineage

Fully open: weights + Dolma pretraining data + training code + training logs.

Sources