Workspace capabilities

Available models

Every language model available on the platform, with its tier and residency.

Four models run in Australia: the three Claude models on AWS Bedrock, and Gemini 2.5 Flash on Google Vertex. Everything else may be served from outside Australia. See data residency.

Each model carries a per-minute rate set by its tier. The model picker shows the current rate beside each model, and Billing covers how per-minute pricing works.

Claude, on AWS Bedrock, Australia

ModelTier
Claude Haiku 4.5 (default)Fast
Claude Sonnet 4.6Balanced
Claude Opus 4.6Power

Gemini, on Google Vertex

ModelTierResidency
Gemini 2.5 FlashFastAustralia
Gemini 2.5 Flash-LiteFastGlobal
Gemini 2.5 ProBalancedGlobal
Gemini 3 Flash PreviewFastGlobal
Gemini 3.1 Flash-LiteFastGlobal
Gemini 3.1 Pro PreviewBalancedGlobal
Gemini 3.5 FlashFastGlobal

OpenAI, global

ModelTier
GPT-4o miniFast
GPT-4.1Balanced
GPT-4.1 nanoFast
GPT-5Balanced
GPT-5 miniFast
GPT-5 nanoFast
GPT-5.1Balanced
GPT-5.2Balanced
GPT-5.4Balanced
GPT-5.4 miniFast
GPT-5.4 nanoFast
GPT-5.5Balanced
o1Power
o3-miniFast
o4-miniFast

Groq, Cerebras and Inception, global

ModelProviderTier
Llama 3.3 70BGroqSpeed
Llama 3.1 8BGroqFast
GPT-OSS 20BGroqFast
GPT-OSS 120BGroqFast
Qwen3.6 27BGroqFast
Gemma 4 31BCerebrasFast
Mercury 2InceptionFast

The list in your workspace is the live source of truth. Models are added and retired, so the picker can differ from the tables above. It always shows the current set, with the default marked.

On this page