AI Universe · Model
Model zoom 2
What is the brain, and how is it made? 5 concept buckets. Each one is a job you can learn once; underneath it, who does that job on each platform.
Concept
Training data
The text, code and images a model learns from, cleaned and de-duplicated.
collectionfilteringde-duplicationlicences
Physical: who does this job, where
| Open source | FineWeb, datatrove |
|---|---|
| Azure | — |
| AWS | — |
| GCP | — |
| Vendors | Scale AI, labelling services |
Concept
Pre-training
Teaching a model language from scratch by predicting the next word, at huge scale.
tokenizertransformerthe training loop
Physical: who does this job, where
| Open source | PyTorch, Megatron, DeepSpeed |
|---|---|
| Azure | Azure ML |
| AWS | SageMaker |
| GCP | Vertex AI, JAX |
| Vendors | frontier labs |
Concept
Fine-tuning and alignment
Adapting a trained model to a task or to preferences.
SFTLoRA / QLoRADPO and RLdistillation
Physical: who does this job, where
| Open source | Hugging Face TRL, Unsloth, PEFT |
|---|---|
| Azure | Foundry fine-tuning |
| AWS | Bedrock custom models |
| GCP | Vertex AI tuning |
| Vendors | OpenAI fine-tuning |
Concept
The models
The trained weights you call: big closed ones, open-weight ones, small ones.
frontieropen-weightsmall and specialisedembedding models
Physical: who does this job, where
| Open source | Llama, Qwen, Mistral, gpt-oss, DeepSeek |
|---|---|
| Azure | Phi; hosts many in Foundry |
| AWS | Nova; hosts many in Bedrock |
| GCP | Gemini, Gemma |
| Vendors | Claude, GPT |
Concept
Model evaluation
How good is the model itself, before it is in any application?
benchmarksleaderboardsred-teaming the model
Physical: who does this job, where
| Open source | lm-evaluation-harness, HELM |
|---|---|
| Azure | Foundry model benchmarks |
| AWS | Bedrock model evaluation |
| GCP | Vertex model evaluation |
| Vendors | LMArena |