fix(openai): correct stale pricing for gpt-4o family and gpt-5.x pro/5.5

[KartikMouli]

Summary

Several models in the OpenAI provider have stale or incorrect pricing in their YAMLs. The most impactful is the gpt-4o ↔ chatgpt-4o-latest swap that has been live since the v1.0.0 plugin migration — anyone using Dify's usage stats for gpt-4o has been billed against double the actual rate. This PR corrects 7 YAMLs against OpenAI's current official pricing.

All values verified against https://developers.openai.com/api/docs/pricing (the current official OpenAI pricing page) and https://platform.openai.com/docs/models.

Pricing corrections

File	Before (in/out)	After (in/out)	Notes
models/openai/models/llm/gpt-4o.yaml	5.00 / 15.00	2.50 / 10.00	Post Aug 6 2024 standard rate; matches Azure plugin's gpt-4o
models/openai/models/llm/chatgpt-4o-latest.yaml	2.50 / 10.00	5.00 / 15.00	chat-latest (ChatGPT model) is priced higher than gpt-4o; values look swapped with above
models/openai/models/llm/gpt-4o-audio-preview.yaml	5.00 / 15.00	2.50 / 10.00	Text I/O matches gpt-4o; audio I/O is metered separately
models/openai/models/llm/gpt-4o-audio-preview-2025-06-03.yaml	5.00 / 15.00	2.50 / 10.00	Same
models/openai/models/llm/gpt-5.5.yaml	5.00 / 22.50	5.00 / 30.00	Output rate per OpenAI pricing page
models/openai/models/llm/gpt-5.5-pro.yaml	15.00 / 90.00	30.00 / 180.00	Pro tier matches gpt-5.4-pro structure
models/openai/models/llm/gpt-5.4-pro.yaml	30.00 / 135.00	30.00 / 180.00	Output rate per OpenAI pricing page

The latter three were introduced in #2981 (merged 2026-04-26) which added the GPT-5.5 family; the values appear to have been estimated before final pricing was published.

Why this matters

gpt-4o is one of the most widely used models on Dify. The current 5.00 / 15.00 configuration causes Dify's cost-estimation UI and token-spend dashboards to display 2× the actual OpenAI bill. The reverse is true for chatgpt-4o-latest (showing 50% of the actual cost). This has been quietly wrong since the v1.0.0 plugin migration (Feb 2025).

This PR contains Changes to Non-Plugin

• Documentation
• Other

This PR contains Changes to Non-LLM Models Plugin

• I have Run Comprehensive Tests Relevant to My Changes

This PR contains Changes to LLM Models Plugin

• My Changes Affect Message Flow Handling (System Messages and User→Assistant Turn-Taking)
• My Changes Affect Tool Interaction Flow (Multi-Round Usage and Output Handling, for both Agent App and Agent Node)
• My Changes Affect Multimodal Input Handling (Images, PDFs, Audio, Video, etc.)
• My Changes Affect Multimodal Output Generation (Images, Audio, Video, etc.)
• My Changes Affect Structured Output Format (JSON, XML, etc.)
• My Changes Affect Token Consumption Metrics
• My Changes Affect Other LLM Functionalities (Reasoning Process, Grounding, Prompt Caching, etc.)
• Other Changes (Add New Models, Fix Model Parameters etc.)

YAML pricing fields only — no runtime/behavior changes. Token consumption metrics shown in the Dify UI and exposed via APIs will now reflect actual OpenAI billing for affected models.

Version Control

• I have Bumped Up the Version in Manifest.yaml (models/openai/manifest.yaml: 0.4.0 → 0.4.1, PATCH for backwards-compatible fix)

Dify Plugin SDK Version

• I have Ensured dify_plugin>=0.3.0,<0.6.0 is in requirements.txt (no SDK changes in this PR)

Environment Verification

SaaS Environment

• Pricing values cross-referenced against OpenAI's official pricing page; YAML-only change with no runtime behavior modified. Following the pattern of the merged precedent fix(gemini): update pricing configuration for Gemini models #2674 (fix(gemini): update pricing configuration for Gemini models) which made an analogous pricing-only YAML correction.

[gemini-code-assist]

Code Review

This pull request updates the OpenAI plugin version to 0.4.1 and adjusts the pricing configuration for several models, including ChatGPT-4o, GPT-4o, and the GPT-5 series. A review comment identifies a missing trailing newline in the gpt-4o-audio-preview-2025-06-03.yaml file, suggesting its addition for POSIX compliance and consistency.

[gemini-code-assist]

The file is missing a trailing newline. Adding a newline at the end of the file ensures consistency with other model configuration files in this repository and adheres to standard POSIX formatting practices.

  currency: USD