Llama-Primus-Merged

Your key comes from /keys. Every request is metered and audited against your account, and the model id is the slug in this page’s address.

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["AD_API_KEY"],
    base_url="https://adversariallm.ai/v1",
)

stream = client.chat.completions.create(
    model="llama-primus-merged",
    messages=[{"role": "user", "content": "…"}],
    stream=True,
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

Llama-Primus-Merged is a cybersecurity-specialized 8B LLM from Trend Micro's AI Lab, built on Llama-3.1-8B-Instruct. Per the card, it was first pre-trained on a large cybersecurity corpus (2.77B tokens), then instruction fine-tuned on approximately 1,000 curated cybersecurity QA tasks, and finally merged back with Llama-3.1-8B-Instruct to maintain the same instruction-following capability. The card reports that this merge preserves general chat ability while achieving a 14.84% improvement in aggregated scores across cybersecurity benchmarks versus the base Llama-3.1-8B-Instruct. The intro frames it in the context of domain-specialized LLMs ("Large Language Models (LLMs) have demonstrated remarkable versatility in recent years, with promising applications in specialized domains such as finance, law, and biomedicine."). The card carries languages English and Japanese and a text-generation pipeline tag, and explicitly notes: "No TrendMicro customer information is included."

from transformers import pipeline

pipe = pipeline("text-generation", model="trendmicro-ailab/Llama-Primus-Merged")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)
```

```python
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("trendmicro-ailab/Llama-Primus-Merged")
model = AutoModelForCausalLM.from_pretrained("trendmicro-ailab/Llama-Primus-Merged", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

No measurements published for this version yet.

Baseline is the strongest general-purpose model we could run on the same suite, same setup, same day. The control row tells you what the other rows are worth.