Built-In Model Pricing and Custom Pricing in CostHQ

CostHQ ships with built-in pricing for 42+ models. Override per-model prices with cs pricing set, or add fine-tuned model pricing namespaced by provider.

CostHQ includes built-in pricing for 42+ AI models, so you never have to calculate costs manually — just provide prompt and completion tokens when logging an API call and CostHQ computes the dollar cost for you.

View current pricing

To see every model price currently active (built-in defaults merged with any custom overrides you have set), run:

cs pricing list

Built-in model families

CostHQ ships with pricing for the following models out of the box.

Anthropic Claude

Model
claude-opus-4
claude-sonnet-4
claude-3-7-sonnet
claude-3-5-sonnet-20241022
claude-3-5-haiku-20241022
claude-3-opus-20240229
claude-3-sonnet-20240229
claude-3-haiku-20240307

OpenAI

Model
gpt-4o
gpt-4o-mini
gpt-4-turbo
gpt-4
gpt-3.5-turbo
o1
o1-mini
o3-mini

OpenAI Codex

Model
gpt-5.1-codex-max
gpt-5.1-codex-mini
gpt-5.3-codex
codex-mini-latest

CostHQ also includes built-in pricing for DeepSeek, Azure OpenAI, and Cohere models (42+ total). Run cs pricing list to see the full table.

Custom pricing

Use cs pricing set to add a new model or override the built-in price for an existing one.

# Add or override a model
cs pricing set my-ft-model 5.00 15.00

# Namespace by provider (prevents collisions)
cs pricing set gpt-4o 2.50 10.00 --provider openai
cs pricing set gpt-4o 3.00 12.00 --provider azure

# Reset to defaults
cs pricing reset

Command parameters

Parameter	Description
<model>	Model name
<in>	Input price per 1 million tokens (USD)
<out>	Output price per 1 million tokens (USD)
--provider <p>	Optional provider namespace — stores the entry as provider/model to prevent name collisions across providers

Where custom pricing is stored

Custom prices are saved to ~/.CostHQ/pricing.json and merged with the built-in defaults at runtime. Deleting this file or running cs pricing reset returns all prices to their built-in values.

Fallback lookup order

When you run cs log-ai, CostHQ resolves the model price in this order:

1. provider/model (namespaced key, e.g. azure/gpt-4o)
2. Plain model name (e.g. gpt-4o)

This means namespaced entries always win over generic ones when both exist.

Unknown models

If you use a model that is not in the pricing table and you have not added a custom price for it, CostHQ cannot calculate the cost automatically. Pass the -c <cost> flag to cs log-ai to provide the dollar amount manually:

cs log-ai -p my-provider -m my-new-model -t 10000 -c 0.12

Pricing source transparency — cs log-ai --json returns a pricing object that tells you exactly where the price came from:

{
  "pricing": {
    "source": "built-in",
    "modelKnown": true,
    "inputPer1M": 3.0,
    "outputPer1M": 15.0
  }
}

source is "built-in", "custom", "local", or "manual". If modelKnown is false, pass -c <cost> explicitly to avoid untracked spend.

Local Model Pricing

For local models (Ollama, vLLM, LM Studio) where costs are based on compute time rather than tokens, use cs local-models add to set an hourly GPU rate, and log usage with cs log-ai --duration. See Local Models for details.