add deepseek-v2, yi-large and gpt-4o pricing to token_counter

[usamimeri]

User description

for usage calculation

---

PR Type

enhancement

---

Description

• Added new model pricing details to the token counter for gpt-4o, gpt-4o-2024-05-13, yi-large, and deepseek-chat.
• Updated the token counting function to support the newly added models.

---

Changes walkthrough 📝

	Relevant files
Enhancement	token_counter.py Add Pricing and Token Counting for New Models                        metagpt/utils/token_counter.py Added pricing details for gpt-4o, gpt-4o-2024-05-13, yi-large, and deepseek-chat models. Updated token counting logic to include new models gpt-4o-2024-05-13 and gpt-4o. +6/-0

---

💡 PR-Agent usage: Comment /help on the PR to get a list of all available PR-Agent tools and their descriptions

[codiumai-pr-agent-pro[bot]]

PR-Agent was enabled for this repository. To continue using it, please link your git user with your CodiumAI identity here.

PR Description updated to latest commit (https://github.com/geekan/MetaGPT/commit/1047abf1f40c6bc460247bba6254c42a6adaa04a)

• [ ] Copy walkthrough table to "Files Changed" Tab

[codiumai-pr-agent-pro[bot]]

PR-Agent was enabled for this repository. To continue using it, please link your git user with your CodiumAI identity here.

PR Review 🔍

⏱️ Estimated effort to review [1-5]	2, because the changes are straightforward and localized to specific sections of the code, involving the addition of new model pricing and updating the token counting logic.
🧪 Relevant tests	No
⚡ Possible issues	Consistency Issue: The pricing for the new models should be verified for consistency with existing pricing strategies and market rates.
	Hardcoding Values: The hardcoded token prices and model names might make the system less flexible and harder to maintain. Consider using a more dynamic approach or external configuration.
🔒 Security concerns	No

[codiumai-pr-agent-pro[bot]]

PR-Agent was enabled for this repository. To continue using it, please link your git user with your CodiumAI identity here.

PR Code Suggestions ✨

Category	Suggestion	Score
Maintainability	Add comments to explain the pricing model for new entries Consider adding a comment to explain the pricing model or the rationale behind the specific pricing values for 'gpt-4o' and 'gpt-4o-2024-05-13'. This will help maintainers understand the pricing decisions and ensure consistency in future updates. metagpt/utils/token_counter.py [40-41] -"gpt-4o": {"prompt": 0.005, "completion": 0.015}, -"gpt-4o-2024-05-13": {"prompt": 0.005, "completion": 0.015}, +"gpt-4o": {"prompt": 0.005, "completion": 0.015}, # Pricing based on XYZ factors +"gpt-4o-2024-05-13": {"prompt": 0.005, "completion": 0.015}, # Pricing based on XYZ factors Suggestion importance[1-10]: 7 Why: Adding comments for clarity on pricing decisions is beneficial for maintainability, though not critical for functionality.	7
Possible issue	Verify the tokenization method for newly added models The addition of 'gpt-4o' and 'gpt-4o-2024-05-13' in the token counting function should be verified for correctness in the context of how tokens are counted. If these models have different tokenization methods, adjustments might be necessary. metagpt/utils/token_counter.py [231-232] -"gpt-4o-2024-05-13", -"gpt-4o", +"gpt-4o-2024-05-13", # Verify tokenization method +"gpt-4o", # Verify tokenization method Suggestion importance[1-10]: 6 Why: Verifying tokenization methods for new models is important to ensure accuracy in token counting, addressing a potential issue in functionality.	6
Best practice	Update the model name to include version or date for consistency Ensure that the model name 'deepseek-chat' is consistent with other model naming conventions in the dictionary. If the model has a version or date, consider including it for clarity and future updates. metagpt/utils/token_counter.py [67] -"deepseek-chat": {"prompt": 0.00014, "completion": 0.00028}, +"deepseek-chat-2024-05-13": {"prompt": 0.00014, "completion": 0.00028}, # Assuming the date is relevant Suggestion importance[1-10]: 5 Why: Ensuring naming consistency is good practice, but the suggestion assumes a date should be added without evidence that it's necessary or standard for all models.	5

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 70.20%. Comparing base (9ff1e2f) to head (b1c1dd2).

:exclamation: Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1274      +/-   ##
==========================================
- Coverage   70.21%   70.20%   -0.01%     
==========================================
  Files         316      316              
  Lines       18866    18866              
==========================================
- Hits        13246    13245       -1     
- Misses       5620     5621       +1     

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.