Skip to content

Models

NucleusIQ uses a provider-agnostic BaseLLM interface. Swap providers without changing the rest of your agent code.

Supported providers

Provider Package Install
OpenAI nucleusiq-openai pip install nucleusiq-openai
Google Gemini nucleusiq-gemini pip install nucleusiq-gemini
Groq nucleusiq-groq pip install nucleusiq-groq
Ollama nucleusiq-ollama pip install nucleusiq-ollama
Mock (testing) Built-in from nucleusiq.core.llms.mock_llm import MockLLM

Usage

from nucleusiq_openai import BaseOpenAI

llm = BaseOpenAI(model_name="gpt-4o-mini")

from nucleusiq_gemini import BaseGemini

llm = BaseGemini(model_name="gemini-2.5-flash")

from nucleusiq_groq import BaseGroq

llm = BaseGroq(model_name="llama-3.3-70b-versatile", async_mode=True)

from nucleusiq_ollama import BaseOllama

llm = BaseOllama(model_name="llama3.2", async_mode=True)

from nucleusiq.core.llms.mock_llm import MockLLM

llm = MockLLM()  # No API key needed

Available models

OpenAI

Model Use case
gpt-4o High capability, multimodal
gpt-4o-mini Fast, cost-effective
gpt-4.1 Latest GPT-4 series
gpt-4.1-mini Balanced performance/cost
gpt-4.1-nano Ultra-fast, lowest cost
o3 Reasoning model
o3-mini Reasoning, cost-effective
o4-mini Latest reasoning model

Gemini

Model Context Thinking
gemini-2.5-pro 1M tokens Yes
gemini-2.5-flash 1M tokens Yes
gemini-2.0-flash 1M tokens No
gemini-1.5-pro 2M tokens No
gemini-1.5-flash 1M tokens No

Groq

Groq rotates Llama, Mixtral, Qwen, GPT-OSS, and other checkpoints frequently — see Groq models. Typical starter IDs:

Model id Typical use
llama-3.3-70b-versatile Chat + local tools
openai/gpt-oss-20b Structured output demos (json_schema)

Beta package nucleusiq-groq 0.1.0b1 requires nucleusiq>=0.7.9 and ships against the official groq Python SDK (>=1.2,<2).

Ollama

Model ids are whatever your Ollama server exposes (ollama list). Typical local tags:

Model id Typical use
llama3.2 General chat + tools (examples default)
mistral, qwen2.5, … Swap names per your catalog

Alpha package nucleusiq-ollama 0.1.0a1 requires nucleusiq>=0.7.10 and uses the official ollama SDK (>=0.5,<1). See Ollama provider guide.

Parameter control

Set LLM parameters at the agent level:

from nucleusiq.agents import Agent
from nucleusiq.agents.config import AgentConfig
from nucleusiq.prompts.zero_shot import ZeroShotPrompt

config = AgentConfig(
    llm_max_output_tokens=1024,
    verbose=True,
)
agent = Agent(
    name="configured-agent",
    prompt=ZeroShotPrompt().configure(system="You are a helpful assistant."),
    llm=llm,  # e.g. BaseOpenAI / BaseGemini / BaseGroq / BaseOllama from the sections above
    config=config,
)

Provider-specific params

Use provider-specific parameter classes for advanced settings:

from nucleusiq_openai import OpenAILLMParams

config = AgentConfig(
    llm_params=OpenAILLMParams(temperature=0.2, reasoning_effort="high"),
)

from nucleusiq_gemini import GeminiLLMParams, GeminiThinkingConfig

config = AgentConfig(
    llm_params=GeminiLLMParams(
        temperature=0.5,
        thinking_config=GeminiThinkingConfig(thinking_budget=2048),
    ),
)

from nucleusiq_groq import GroqLLMParams

config = AgentConfig(
    llm_params=GroqLLMParams(
        temperature=0.5,
        parallel_tool_calls=True,
    ),
)

from nucleusiq_ollama import OllamaLLMParams

config = AgentConfig(
    llm_params=OllamaLLMParams(
        temperature=0.5,
        think="high",
        keep_alive="10m",
    ),
)

Per-task overrides

Override parameters for a single execution:

from nucleusiq_openai import OpenAILLMParams

result = await agent.execute(
    task,
    llm_params=OpenAILLMParams(temperature=0.0),
)

from nucleusiq_gemini import GeminiLLMParams

result = await agent.execute(
    task,
    llm_params=GeminiLLMParams(temperature=0.0),
)

from nucleusiq_groq import GroqLLMParams

result = await agent.execute(
    task,
    llm_params=GroqLLMParams(temperature=0.0, max_output_tokens=512),
)

from nucleusiq_ollama import OllamaLLMParams

result = await agent.execute(
    task,
    llm_params=OllamaLLMParams(temperature=0.0, max_output_tokens=512),
)

See also