✨ Change OpenAI models to better and cheaper options

[jernejk]

Cc:

Hi @bradystroud and @adamcogan,

Pain

The default model and the options we give users are currently not the best options for neither performance nor cost.

This feature should be tackled soon for 3 reasons:

1. GPT-4o-mini is a lot better and cheaper than GPT-3.5-Turbo (current default model)
2. GPT-3.5 and GPT-3.5-Turbo are slowly being phased out
3. First impressions are usually very important. While GPT-3.5 is great at well defined tasks and well formed data, it can quickly start to fall apart when introducing complexity.

Moving from GPT-3.5-Turbo to 4o-mini will likely save us $0.90 / 1M tokens (most of the cost are output tokens) and we'll have much more accurate responses as well!

Suggested Solution

Remove model GPT-3.5 and use GPT-4o-mini instead.

Pricing (sorted by cost)

Model	Input	Output	Is an good option
gpt-4o-mini	$0.15 / 1M	tokens $0.60 / 1M tokens	✅ (as default)
gpt-3.5-turbo-0125	$0.50 / 1M tokens	$1.50 / 1M tokens	❌
gpt-3.5-turbo-0301	$1.50 / 1M tokens	$2.00 / 1M tokens	❌
gpt-4o	$2.50 / 1M	tokens $10.00 / 1M tokens	✅
o1-mini	$3.00 / 1M	tokens $12.00 / 1M tokens	✅
o1	$15.00 / 1M	tokens $60.00 / 1M tokens	✅
gpt-4	$30.00 / 1M tokens	$60.00 / 1M tokens	❌
gpt-3.5	n/a*	n/a*	❌

Figure: Official pricing on 09/01/2025 - https://openai.com/api/pricing/

* GPT-3.5 is no longer listed for pricing.

Tasks

• Remove 3.5-Turbo and 4.0 models
• Add 4o-mini, 4o, o1 and o1-mini models (o1 is still cheaper than current gpt-4 model used in RulesGPT!)

Acceptance Criteria

Users should be able to use the latest and greatest models at lower cost and higher accuracy.

Thanks!