Repository navigation
docs(leaderboard): add DeepSeek results - #286
wilburx813 wants to merge 3 commits into
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b99793d5af
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| {"id": "robotwin--gpt-5-6-no-reasoning", "benchmark_id": "robotwin", "configuration_id": "gpt-5-6-no-reasoning", "mean_elapsed_seconds": 957.4, "total_output_tokens": 2456401}, | ||
| {"id": "robotwin--gpt-6-astra-low", "benchmark_id": "robotwin", "configuration_id": "gpt-6-astra-low", "mean_elapsed_seconds": 1078.8, "total_output_tokens": 2763641} | ||
| {"id": "robotwin--gpt-6-astra-low", "benchmark_id": "robotwin", "configuration_id": "gpt-6-astra-low", "mean_elapsed_seconds": 1078.8, "total_output_tokens": 2763641}, | ||
| {"id": "libero-pro--deepseek-v4-1-flash-xhigh", "benchmark_id": "libero-pro", "configuration_id": "deepseek-v4-1-flash-xhigh", "mean_elapsed_seconds": 963.25, "source_ids": ["deepseek-libero-pro-800-20261001"], "total_output_tokens": 5229711.52} |
There was a problem hiding this comment.
On the Time & Token Costs page, this key makes the renderers label 5,229,712 as “Total output tokens” (docs/_ext/leaderboard.py:41-42,293-296), but the new source note at line 883 says the value averages total_input_tokens + total_output_tokens across 800 episodes; the fractional value also cannot be a literal token total. This makes DeepSeek incomparable with adjacent output-only totals and misstates both the token scope and aggregation, so retain the supported mean_total_tokens field or replace the value with the actual integer output-token sum.
AGENTS.md reference: AGENTS.md:L5-L6
Useful? React with 👍 / 👎.
1 participant
Add DeepSeek V4.1 results to the LIBERO-PRO leaderboard.