Skip to content

Available Models

uniGPT offers a variety of models — both internal (hosted on-premises at the university) and external (provided via cloud partners).

For balance reset details and practical advice on conserving credits, see Usage Limits & Balance.

Internal Models (On-Premises)

These models run entirely on university infrastructure. Your data never leaves the university, making them fully GDPR-compliant and suitable for sensitive data.

Model Description Input / output cost (credits)
Mistral Small Fast, efficient general-purpose model 0.15 / 0.2
Llama-3.3-70B Older open-weight model by Meta 0.65 / 0.5
gpt-oss-120b Large internal model for complex tasks 0.15 / 0.6
gemma-3-27b-it Efficient multimodal model with image support 0.09 / 0.16
medgemma-1.5-4b-it Specialized multimodal model for medical text and image understanding 0.02 / 0.04
Qwen3.5-35B-A3B Chinese Advanced reasoning and coding model with image support 0.035 / 0.138

Best for sensitive data

Use internal models when working with personal data, unpublished research, or any data subject to GDPR restrictions.

External Models (Cloud Providers)

These models are accessed via cloud providers through the GÉANT OCRE framework. While they offer state-of-the-art performance, data is processed externally.

Model Provider Notes Input / output cost (credits)
GPT-5.4 OpenAI (via Azure) Strong flagship reasoning model 2.5 / 15
GPT-5.2 OpenAI (via Azure) High-capability reasoning model 1.75 / 14
GPT-5 Mini OpenAI (via Azure) Lower-cost GPT-5 variant 0.25 / 2
Gemini 3.1 Pro Google Cloud Advanced multimodal model 2 / 12
Gemini 3 Flash Google Cloud Fast multimodal model 0.25 / 1.5
Claude Opus 4.6 Anthropic (via Google) Strongest Claude model for deep analysis 5 / 25
Claude Sonnet 4.5 Anthropic (via Google) Balanced writing and analysis model 3 / 15
Claude Haiku 4.5 Anthropic (via Google) Fastest Claude option 1 / 5

Cost awareness

External models cost significantly more per token than internal models. The university receives only a minimal discount (<5%) from providers. Please use external models sparingly and prefer internal models or cost-effective options like Gemini (Flash) when possible. Otherwise, user limits may need to be reduced.

Choosing the Right Model

  • Quick tasks & drafting: Mistral Small, Llama-3.3-70B
  • Complex reasoning: gpt-oss-120b, GPT-5, Claude Opus
  • Medical text & image tasks: medgemma-1.5-4b-it
  • Cost-effective external option: Gemini Pro
  • Sensitive/confidential data: Always use internal models

For current balance behavior and reset timing, see Usage Limits & Balance.