Model results

Video-MME-v2

Choose a shared evaluation protocol to compare model scores. Versions, sources and configurations stay separate.

Atlas snapshot captured 10/19/2026, 10:20:28 AM UTC. This is an import timestamp, not the evaluation date. Coverage may be incomplete.

Reviewed recommendation and limits

Use grouped questions to examine whether video understanding stays consistent across related observations and reasoning steps.

Keep grouped nonlinear score separate from per-question accuracy. Match dependency scoring, sampled frames, subtitle alignment, audio input and reasoning mode.. Imported groups below may use different protocols; this review does not certify each run. Separate subtitle text from raw audio, frame budgets and thinking settings. Public data does not prove freedom from contamination. Dataset use is restricted to academic research.

Publisher results Video-MME-v2 methodology and leaderboard Video-MME-v2 paper Video-MME-v2 evaluation and use restrictions

← All evaluation groups

Video-MME-v2 · video-mme-v2

Source: Hugging Face eval result
Version: Not reported
Measurement date: 2026-04-30
Metric: % · Higher is better
Evaluation scope: Registered model evaluation · VideoMMEv2 leaderboard

Positions are withheld because at least one result lacks explicit independent provenance or a documented evaluation scope. Missing metadata is shown as unreported; matching imported metadata is not independent verification of the protocol.

Kimi K2.5
moonshotai/kimi-k2.5-0127
61.1%
View modelEvidence Evidence class unreported
Qwen3.5 397B A17B
qwen/qwen3.5-397b-a17b-20260216
55.9%
View modelEvidence Evidence class unreported