long context benchmark
LongVideoBench Leaderboard
LongVideoBench is a question-answering benchmark with video-language interleaved inputs up to an hour long.
Updated Sep 5, 2026
Models4
Model coverage4
MetricScore
EvidenceB
LongVideoBench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelBY | Score80.60% | Percentile100.00% | Participants4 | EvidenceC | Evaluated |
| Rank02 | ModelBY | Score80.60% | Percentile66.67% | Participants4 | EvidenceC | Evaluated |
| Rank03 | ModelMA | Score79.80% | Percentile33.33% | Participants4 | EvidenceC | Evaluated |
| Rank04 | ModelAC | Score54.70% | Percentile0.00% | Participants4 | EvidenceC | Evaluated |
LongVideoBench Highlights
The leading models and scores on this benchmark.
LongVideoBench Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for LongVideoBench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is LongVideoBench?
What LongVideoBench measures and how its scores work.
It consists of 3,763 web-collected videos with subtitles and 6,678 human-annotated multiple-choice questions across 17 fine-grained categories.
It measures long-term multimodal understanding using Score, reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- LongVideoBench
- Modality
- multimodal
- Primary category
- long context
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about LongVideoBench.
Which model scores highest on LongVideoBench?
Seed 2.1 Pro is currently ranked first with 80.60%.
What are the top three models on LongVideoBench?
The current leaders are Seed 2.1 Pro (80.60%), Seed 2.1 Turbo (80.60%), and Kimi K2.5 (79.80%).
Which LongVideoBench model has the lowest official input price?
Qwen2.5 VL 7B Instruct has the lowest matched official input price at $0.35 input / $1.1 output per 1M tokens.
Which models are fastest among LongVideoBench results?
No matched runtime record is currently available.
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.