Supported Models

Every model in this list works out of the box with Squadron. Add your API key to a model block and reference any of the keys below as models.<config_name>.<key>. See Models for configuration details.

Prices are per 1M tokens, last verified April 2026. Use pricing blocks on your model config to override them.

Anthropic


model "anthropic" {
  provider = "anthropic"
  api_key  = vars.anthropic_api_key
}
 
# Reference: models.anthropic.claude_sonnet_4_6

Key	API name	Input	Output	Cache read	Cache write
`claude_opus_4_6`	`claude-opus-4-6`	$5.00	$25.00	$0.50	$6.25
`claude_sonnet_4_6`	`claude-sonnet-4-6`	$3.00	$15.00	$0.30	$3.75
`claude_haiku_4_5`	`claude-haiku-4-5-20251001`	$1.00	$5.00	$0.10	$1.25
`claude_opus_4`	`claude-opus-4-20250514`	$15.00	$75.00	$1.50	$18.75
`claude_sonnet_4`	`claude-sonnet-4-20250514`	$3.00	$15.00	$0.30	$3.75
`claude_3_5_sonnet`	`claude-3-5-sonnet-20241022`	$3.00	$15.00	$0.30	$3.75
`claude_3_5_haiku`	`claude-3-5-haiku-20241022`	$0.80	$4.00	$0.08	$1.00

OpenAI


model "openai" {
  provider = "openai"
  api_key  = vars.openai_api_key
}
 
# Reference: models.openai.gpt_5

Key	API name	Input	Output	Cache read	Cache write
`gpt_5`	`gpt-5`	$2.50	$15.00	$0.25	$2.50
`gpt_5_mini`	`gpt-5-mini`	$0.75	$4.50	$0.075	$0.75
`gpt_5_nano`	`gpt-5-nano`	$0.20	$1.25	$0.02	$0.20
`gpt_4_1`	`gpt-4.1`	$2.00	$8.00	$0.50	$2.00
`gpt_4_1_mini`	`gpt-4.1-mini`	$0.40	$1.60	$0.10	$0.40
`gpt_4_1_nano`	`gpt-4.1-nano`	$0.10	$0.40	$0.025	$0.10
`gpt_4o`	`gpt-4o`	$2.50	$10.00	$1.25	$2.50
`gpt_4o_mini`	`gpt-4o-mini`	$0.15	$0.60	$0.075	$0.15
`gpt_4_turbo`	`gpt-4-turbo`	$10.00	$30.00	$10.00	$10.00
`o1`	`o1`	$15.00	$60.00	$7.50	$15.00
`o1_mini`	`o1-mini`	$1.10	$4.40	$0.55	$1.10
`o3`	`o3`	$10.00	$40.00	$2.50	$10.00
`o3_mini`	`o3-mini`	$1.10	$4.40	$0.55	$1.10
`o4_mini`	`o4-mini`	$1.10	$4.40	$0.55	$1.10

Google Gemini


model "gemini" {
  provider = "gemini"
  api_key  = vars.gemini_api_key
}
 
# Reference: models.gemini.gemini_2_5_pro

Key	API name	Input	Output	Cache read	Cache write
`gemini_2_5_pro`	`gemini-2.5-pro`	$1.25	$10.00	$0.125	$1.25
`gemini_2_5_flash`	`gemini-2.5-flash`	$0.30	$2.50	$0.03	$0.30
`gemini_2_5_flash_lite`	`gemini-2.5-flash-lite`	$0.10	$0.40	$0.01	$0.10
`gemini_2_0_flash`	`gemini-2.0-flash`	$0.10	$0.40	$0.025	$0.10
`gemini_2_0_flash_exp`	`gemini-2.0-flash-exp`	$0.10	$0.40	$0.025	$0.10
`gemini_1_5_pro`	`gemini-1.5-pro`	$1.25	$5.00	$0.3125	$1.25
`gemini_1_5_flash`	`gemini-1.5-flash`	$0.075	$0.30	$0.01875	$0.075

Ollama (local models)

The Ollama provider connects to any OpenAI-compatible local inference server — Ollama itself, vLLM, llama.cpp, LM Studio, and others. Because Squadron can’t know what models you have installed, you define them with aliases instead of picking from a built-in list.


model "local" {
  provider = "ollama"
  base_url = "http://localhost:11434/v1"
  aliases = {
    gemma4     = "gemma4"
    gemma4_26b = "gemma4:26b"
    nemotron   = "nemotron-cascade-2:30b"
  }
}
 
# Reference: models.local.gemma4_26b

Token usage is tracked for every turn, but dollar cost is $0 — you’re paying for the hardware, not the tokens.

Don’t see the model you want?

Cloud providers (Anthropic, OpenAI, Gemini) support custom aliases — you can point any HCL-friendly key at any API model name the provider exposes, even ones that aren’t in the tables above.
For brand new releases that don’t yet have built-in pricing, add a pricing override on your model block and Squadron will use your numbers for cost tracking.