Architecture
data/models.yaml. Every label is auditable
against the model's sources.
Specs
- Architecture
- dense
- Total params
- 32B
- Active params
- 32B
- Context window
- 66K tokens
- Attention
- gqa
- Position encoding
- rope
- Pretraining tokens
- 6.0T
- Post-training
- sft, dpo, rlvr
- 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 | 85.4 | as of 2025-11-20 | source ↗ |
Held-out / arena
| IFEval | 89.0 | as of 2025-11-20 | source ↗ |
Available quantizations
GGUF llama.cpp's container; the common local format, k-quants from Q2 to Q8.
runs on llama.cpp, Ollama
FP8 8-bit float, frequently a native release on Hopper / Blackwell GPUs.
runs on vLLM, SGLang, TensorRT-LLM
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
- · [object Object]
- · Four variant tracks (Base, Think, Instruct, RL-Zero)
- · RL-Zero pathway openly released
- · Dolma 3 and Dolci stacks public
Lineage
AI2's first fully open 32B frontier model; built on OLMo 3-Base, a new pretrain in the OLMo 3 family rather than a fine-tune of OLMo 2.