This is a snapshot of popular LLMs available or being rolled out on the Hatz AI platform. Model availability can vary by organization, package, tenant policy, feature flag, and controlled rollout status. If you see a model listed here but do not see it in your model selector, it may be disabled for your organization or still in a controlled rollout period.
A model's lab or company of origin is separate from where inference runs when you use that model through Hatz. Hatz does not route model inference to non-US lab public APIs. Some models run on Hatz-managed cloud infrastructure, and others may run through trusted model providers with contractual or operational retention controls. For exact model-provider and retention details for your organization, contact Hatz Security.
Amazon
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
Amazon Nova 2 Lite | Standard | Medium | Fast | Y | Y | Fast, affordable reasoning with document understanding |
Amazon Nova Lite | Standard | Medium | Fast | Y | N | Fast and affordable with vision capabilities |
Amazon Nova Micro | Standard | Low | Fast | Y | N | Amazon's fastest model for quick tasks |
Amazon Nova Pro | Standard | High | Medium | Y | N | Balanced Amazon model for general tasks |
Anthropic
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
Claude Haiku 4.5 | Standard | Medium | Fast | Y | N | Near-instant replies for simple questions |
Claude 3 Haiku | Standard | Medium | Fast | Y | N | Fast and affordable for simple tasks |
Claude 3 Sonnet | Premium | High | Medium | Y | N | Previous-generation general purpose model |
Claude 4 Opus | Premium | High | Medium | Y | Y | Previous-generation reasoning model |
Claude 4 Sonnet | Premium | High | Medium | Y | Y | Balanced model for general tasks |
Claude 4.5 Opus | Premium | High | Medium | Y | Y | Advanced reasoning and coding capabilities |
Claude 4.5 Sonnet | Premium | High | Medium | Y | Y | Smart and reliable for everyday tasks |
Claude Opus 4.6 | Premium | High | Medium | Y | Y | Superseded by Claude Opus 4.7. |
Claude Opus 4.7 | Premium | High | Medium | Y | Y | Anthropic's most intelligent model with exceptional coding, reasoning, and agentic capabilities. Features a 1M context window, 128K max output, and adaptive thinking. |
Claude Sonnet 4.6 | Premium | High | Fast | Y | Y | Smart and reliable for everyday tasks |
DeepSeek
Security note: When these model families are offered in Hatz, inference is not sent to the model lab's public non-US API. Hatz uses managed infrastructure or trusted model providers and applies provider-specific retention controls. Availability and retention specifics can vary by model and organization, so contact Hatz Security for formal details.
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
DeepSeek V3.2 | Standard | High | Medium | Y | Y | Agentic reasoning for complex multi-step tasks |
Deepseek R1 | Standard | High | Slow | N | Y | Advanced reasoning for math and logic problems |
DeepSeek V4 Pro | Controlled rollout | High | Medium | Y | Y | Advanced reasoning and coding model in controlled rollout |
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
Gemini 2.0 Flash | Standard | Medium | Fast | N | N | Previous-generation fast Google model |
Gemini 2.0 Flash Lite | Standard | High | Fast | N | N | Lightweight Google model for quick tasks |
Gemini 2.5 Flash | Standard | Medium | Fast | Y | N | Lightning fast responses for quick answers |
Gemini 2.5 Flash Image (Nano Banana) | Standard | Medium | Fast | N | N | Generates images from text descriptions |
Gemini 3 Flash | Standard | Medium | Fast | Y | Y | Fast and efficient with thinking mode |
Gemini 3 Pro Preview | Standard | High | Medium | Y | Y | Google's most intelligent model for complex tasks |
Gemini 3.1 Pro Preview | Standard | High | Medium | Y | Y | Google's latest and most capable model |
Gemini 2.5 Pro | Premium | High | Fast | Y | Y | Advanced reasoning and analysis capabilities |
Gemini 3 Pro Image (Nano Banana Pro) | Premium | High | Medium | N | N | Creates and edits images from text descriptions |
Gemini 3.1 Flash Image (Nano Banana 2) | Premium | High | Medium | N | N | Fast image generation with editing support |
Meta
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
Llama 3.1 8B Instruct | Standard | Medium | Medium | N | N | Lightweight open-source model |
Llama 3.1 70B Instruct | Standard | Medium | Fast | N | N | Previous-generation Llama model |
Llama 3.2 1B Instruct | Standard | Medium | Fast | N | N | Ultra-compact model for simple tasks |
Llama 3.2 11B Instruct | Standard | High | Slow | N | N | Understands and describes images |
Llama 4 Maverick 17B Instruct | Standard | High | Fast | Y | N | Free, open-source model with vision capabilities |
Llama 4 Scout 17B Instruct | Premium | High | Fast | Y | N | Free, open-source option for general tasks |
MiniMax
Security note: When these model families are offered in Hatz, inference is not sent to the model lab's public non-US API. Hatz uses managed infrastructure or trusted model providers and applies provider-specific retention controls. Availability and retention specifics can vary by model and organization, so contact Hatz Security for formal details.
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
MiniMax M2.5 | Standard | High | Medium | Y | N | Agent-native model for tool-heavy workflows |
Mistral
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
Mistral 7B Instruct | Standard | Low | Medium | N | N | Lightweight model for basic tasks |
Mistral 8x7B Instruct | Standard | Medium | Fast | N | N | Efficient open-weight model for general use |
Mistral Large | Premium | High | Medium | N | N | Previous-generation Mistral flagship |
Mistral Large 3 | Premium | High | Medium | N | N | Fluent in multiple languages |
Moonshot AI
Security note: When these model families are offered in Hatz, inference is not sent to the model lab's public non-US API. Hatz uses managed infrastructure or trusted model providers and applies provider-specific retention controls. Availability and retention specifics can vary by model and organization, so contact Hatz Security for formal details.
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
Kimi K2 Thinking | Standard | High | Slow | Y | Y | Deep thinking model for complex reasoning |
Kimi K2.5 | Standard | High | Medium | Y | N | Multimodal reasoning with image understanding |
Kimi K2.6 | Controlled rollout | High | Medium | Y | N | Long-context multimodal assistance in controlled rollout |
Kimi K2.7 Code | Controlled rollout | High | Medium | Y | N | Coding-focused Kimi model in controlled rollout |
NVIDIA
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
NVIDIA Nemotron 3 Super 120B A12B | Standard | High | Fast | Y | N | Long-context model for extensive documents |
NVIDIA Nemotron 3 Ultra | Controlled rollout | High | Medium | Y | N | Large open model for enterprise agent and document workflows |
OpenAI
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
GPT 3.5 Turbo | Standard | Medium | Fast | Y | N | Fast and affordable for basic tasks |
GPT 4o | Standard | High | Medium | Y | N | Previous-generation flagship with image support |
GPT-4.1 | Standard | High | Medium | Y | N | Previous-generation model with image capabilities |
GPT-4.1 Mini | Standard | Medium | Fast | Y | N | Previous-generation fast and efficient |
GPT-4.1 Nano | Standard | Medium | Fast | Y | N | Previous-generation lightweight model |
GPT-5 | Standard | High | Medium | Y | Y | Previous-generation model with broad capabilities |
GPT-5 Chat | Standard | High | Medium | Y | N | Previous-generation chat model |
GPT-5 Mini | Standard | Medium | Fast | Y | Y | Fast and efficient for everyday tasks |
GPT-5 Nano | Standard | Medium | Fast | Y | Y | Ultra-lightweight for quick interactions |
GPT-5.1 | Standard | High | Medium | Y | Y | Previous-generation model with reasoning |
GPT-5.4 Mini | Standard | Medium | Fast | Y | N | OpenAI's fast model for everyday tasks |
GPT-5.4 Nano | Standard | Medium | Fast | Y | N | Ultra-fast lightweight model for quick responses |
o3 mini | Standard | Medium | Medium | Y | Y | Previous-generation efficient reasoning |
o4 mini | Standard | High | Medium | Y | Y | Fast reasoning for coding and math |
GPT 4 | Premium | High | Medium | Y | N | Previous-generation flagship model |
GPT-5.2 | Premium | High | Medium | Y | Y | Great for writing emails, essays, and general questions |
GPT-5.4 | Premium | High | Medium | Y | Y | OpenAI's most capable model for professional work |
o1 | Premium | High | Slow | Y | Y | Original OpenAI reasoning model |
o3 | Premium | High | Slow | Y | Y | Best for coding and technical problems |
GPT OSS 120B | Controlled rollout | High | Medium | Y | N | Open-weight model option in controlled rollout |
Qwen
Security note: When these model families are offered in Hatz, inference is not sent to the model lab's public non-US API. Hatz uses managed infrastructure or trusted model providers and applies provider-specific retention controls. Availability and retention specifics can vary by model and organization, so contact Hatz Security for formal details.
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
Qwen3 Coder Next | Standard | High | Medium | Y | N | Optimized for code generation and debugging |
xAI
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
Grok 2 | Standard | Medium | Medium | Y | N | xAI's previous-generation model |
Grok 2 Vision | Standard | Medium | Medium | Y | N | Previous-generation vision model from xAI |
Grok 3 Mini | Standard | Medium | Fast | Y | N | Fast and affordable for general tasks |
Grok 3 | Premium | High | Medium | Y | N | Previous-generation high-performance model |
Grok 3 Fast | Premium | Medium | Fast | Y | N | Previous-generation fast model from xAI |
Grok 3 Mini Fast | Premium | Medium | Fast | Y | N | xAI's fastest model for instant replies |
Grok 4 | Premium | High | Medium | Y | Y | xAI's most powerful model for complex tasks |
Z AI
Security note: When these model families are offered in Hatz, inference is not sent to the model lab's public non-US API. Hatz uses managed infrastructure or trusted model providers and applies provider-specific retention controls. Availability and retention specifics can vary by model and organization, so contact Hatz Security for formal details.
LLM Name | Tier | Intelligence | Speed | Tools (Y/N) | Reasoning (Y/N) | What is it good at? |
GLM 5 | Standard | High | Medium | Y | N | General-purpose frontier model |
GLM 5.1 | Controlled rollout | High | Medium | Y | N | Multilingual and structured-output model in controlled rollout |
GLM 5.2 | Controlled rollout | High | Medium | Y | N | Updated GLM model for multilingual and agentic tasks in controlled rollout |
Some LLM Terminology:
Token
The smallest unit of data that an LLM processes
Usually about 3/4 of a word
If you want to see how many INPUT tokens something is, use this site (it works best for OpenAI models):
***In some cases, tokens are not just words. If you upload a PDF with images, tokens will applied to interpret the images in addition to the words. If you upload a PNG or JPG with no words, tokens will still be used to understand the image.
Context Window
The segment of text that an AI model considers at any given moment when processing or generating language
You can think of this as the size of the input
If a file or prompt is very large or dense, the selected model may not be able to use all of it at once. Try splitting the file, asking for a summary first, starting a new chat for a distinct task, or choosing a model with stronger long-context performance.
Knowledge Cutoff Date
The date that the LLM stopped "learning" new information. If you ask GPT 3.5 Turbo about an event that happened yesterday, it will not have that data
Reasoning
Reasoning models are a type of AI model that are trained to engage in "thinking" before providing an answer, breaking down problems into steps and showing their reasoning process, unlike models that directly generate answers.
They can take more time to answer because they are reflecting on their analysis step-by-step.
Vision
Computer vision gives LLMs the ability to "see" and interpret visual data like images and videos, enabling them to identify objects, classify scenes, and perform tasks that mimic human vision.
More Info
To dive deeper into these LLMs, check out these resources*
*Some features outlined on these sites might not be available on the Hatz AI platform, or they might work differently than explained here. Use these articles to better understand the token limits, reasoning models, and use cases.