Skip to content

feat: add OpenAI o1 and GPT-4o mini models#2003

Open
KushalLukhi wants to merge 1 commit intomckaywrigley:mainfrom
KushalLukhi:feat/add-openai-o1-models
Open

feat: add OpenAI o1 and GPT-4o mini models#2003
KushalLukhi wants to merge 1 commit intomckaywrigley:mainfrom
KushalLukhi:feat/add-openai-o1-models

Conversation

@KushalLukhi
Copy link

Description

Adds support for OpenAI's latest model releases: o1-preview, o1-mini, and GPT-4o mini.

New Models

o1-preview

  • Advanced reasoning model designed for complex tasks
  • Pricing: $15/1M input, $60/1M output tokens
  • No vision support (reasoning-focused)

o1-mini

  • Cost-effective reasoning model, faster than o1-preview
  • 80% cheaper than o1-preview
  • Pricing: $3/1M input, $12/1M output tokens
  • No vision support

GPT-4o mini

  • Most cost-efficient GPT-4o model
  • Replaces GPT-3.5 Turbo for most use cases
  • Vision capabilities included
  • Pricing: $0.15/1M input, $0.60/1M output tokens
  • 60% cheaper than GPT-3.5 Turbo with better performance

Changes

  • Updated lib/models/llm/openai-llm-list.ts
  • Updated types/llms.ts with new model IDs

Testing

  • Models appear in model selection dropdown
  • API requests route correctly
  • Pricing displays accurately
  • Vision capabilities correctly flagged

Future Work

Consider making o1 models visually distinct in UI (they use different system prompt handling).


Closes feature gap with OpenAI's latest API capabilities.

Adds support for OpenAI's newest model releases:

- o1-preview: Advanced reasoning model (5/0 per 1M)
- o1-mini: Fast, cost-effective reasoning (/2 per 1M)
- GPT-4o mini: GPT-4o optimized for efficiency (/bin/bash.15//bin/bash.60 per 1M)

Models added:
- o1-preview - high-quality reasoning
- o1-mini - faster, cheaper reasoning
- GPT-4o mini - better than GPT-3.5 at 60% lower cost

Note: o1 models do not support image input currently.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant