Detail	Value
Base	OLMo-2-0325-32B-Instruct
Parameters	32B
Precision	BF16
Context Length	Up to 65K tokens
License	Apache 2.0
Architecture	OLMo-2 (transformer)

API Usage

Via OpenRouter (Recommended — Cheapest)

python

import requests

response = requests.post(
    "https://openrouter.ai/api/v1/chat/completions",
    headers={
        "Authorization": "Bearer YOUR_OPENROUTER_KEY",
        "Content-Type": "application/json"
    },
    json={
        "model": "allenai/olmo-2-0325-32b-instruct",
        "messages": [
            {"role": "user", "content": "Hello, who are you?"}
        ]
    }
)
print(response.json()["choices"][0]["message"]["content"])

Via DeepInfra

python

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_DEEPINFRA_KEY",
    base_url="https://api.deepinfra.com/v1/openai"
)

response = client.chat.completions.create(
    model="allenai/OLMo-2-0325-32B-Instruct",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

Self-Hosted (vLLM)

bash

pip install vllm
vllm serve allenai/OLMo-2-0325-32B-Instruct --tensor-parallel-size 2

Performance

OLMo-2-32B is the first fully open model to outperform GPT-3.5-Turbo and GPT-4o-mini on key benchmarks.

Benchmark	Score
GSM8K	86.6
MATH	42.0
IFEval	80.3
Hellaswag	86.2
ARC-C	65.1

Citation

bibtex

@article{olmo2025,
  title={OLMo 2: The Best Fully Open Language Model to Date},
  author={Ai2},
  year={2025}
}

fuhaddesmond/fuhad-32b-instruct

🧠 Fuhad-32B-Instruct

Why "Truly Open Source"?

Model Details

API Usage

Via OpenRouter (Recommended — Cheapest)

Via DeepInfra

Self-Hosted (vLLM)

Performance

Citation

fuhaddesmond/
fuhad-32b-instruct