Skip to content

Add MiMo V2.5 Pro (Xiaomi) to the board#7

Closed
lakshvantb wants to merge 1 commit into
mainfrom
board/mimo-v2.5-pro-20260722-1400
Closed

Add MiMo V2.5 Pro (Xiaomi) to the board#7
lakshvantb wants to merge 1 commit into
mainfrom
board/mimo-v2.5-pro-20260722-1400

Conversation

@lakshvantb

Copy link
Copy Markdown
Collaborator

Board rows for mimo-v2.5-pro, generated with generate_model_rows.py --append (the PR #6 mode) from the eval-run data on eval/mimo-v2.5-pro-20260721-1654 — this is the live validation of the automated board-PR path (livebench-private #172).

category score
math 76.0
coding 68.1
data_analysis 48.7
instruction_following 39.7
language 62.9
reasoning 66.0
agentic_coding_v2 41.4
overall 57.5

Cost (Xiaomi first-party pricing 0.435/0.0036/0.87 per M): $0.0166/question, $0.0289/successful task — cheapest model on the board.

Verified: generator self-verify (all columns), check_data.js 43/43 with cost, no orphans, full 1270-question coverage, modelLinks entry (openweight+reasoner).

Merge after the data PR (livebench-private #170); deploy manually as usual.

🤖 Generated with Claude Code

@lakshvantb

Copy link
Copy Markdown
Collaborator Author

Superseded by #8: the workflow E2E run retried the original run's failed questions and recovered answers, so #8's rows come from strictly better data (this PR was generated from the older eval-branch snapshot).

@lakshvantb lakshvantb closed this Jul 22, 2026
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