long context benchmark
LVBench Leaderboard
LVBench evaluates multimodal models' understanding of videos up to two hours in duration across 6 major categories and 21 subcategories.
Updated Sep 5, 2026
Models28
Model coverage28
MetricScore
EvidenceB
LVBench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelGO | Score87.10% | Percentile100.00% | Participants28 | EvidenceC | Evaluated |
| Rank02 | ModelGO | Score85.40% | Percentile96.30% | Participants28 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score81.80% | Percentile92.59% | Participants28 | EvidenceC | Evaluated |
| Rank04 | ModelBY | Score78.00% | Percentile88.89% | Participants28 | EvidenceC | Evaluated |
| Rank05 | ModelBY | Score76.80% | Percentile85.19% | Participants28 | EvidenceC | Evaluated |
| Rank06 | ModelAC | Score76.60% | Percentile81.48% | Participants28 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Score76.60% | Percentile77.78% | Participants28 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score76.20% | Percentile74.07% | Participants28 | EvidenceC | Evaluated |
| Rank09 | ModelMA | Score75.90% | Percentile70.37% | Participants28 | EvidenceC | Evaluated |
| Rank10 | ModelAC | Score74.40% | Percentile66.67% | Participants28 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Score73.60% | Percentile62.96% | Participants28 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score71.40% | Percentile59.26% | Participants28 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Score71.40% | Percentile55.56% | Participants28 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Score67.70% | Percentile51.85% | Participants28 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score63.80% | Percentile48.15% | Participants28 | EvidenceC | Evaluated |
| Rank16 | ModelAC | Score63.60% | Percentile44.44% | Participants28 | EvidenceC | Evaluated |
| Rank17 | ModelAC | Score62.60% | Percentile40.74% | Participants28 | EvidenceC | Evaluated |
| Rank18 | ModelAC | Score62.50% | Percentile37.04% | Participants28 | EvidenceC | Evaluated |
| Rank19 | ModelAC | Score59.20% | Percentile33.33% | Participants28 | EvidenceC | Evaluated |
| Rank20 | ModelAC | Score58.00% | Percentile29.63% | Participants28 | EvidenceC | Evaluated |
| Rank21 | ModelAC | Score56.20% | Percentile25.93% | Participants28 | EvidenceC | Evaluated |
| Rank22 | ModelAC | Score55.80% | Percentile22.22% | Participants28 | EvidenceC | Evaluated |
| Rank23 | ModelAC | Score53.50% | Percentile18.52% | Participants28 | EvidenceC | Evaluated |
| Rank24 | ModelAC | Score49.00% | Percentile14.81% | Participants28 | EvidenceC | Evaluated |
| Rank25 | ModelAC | Score47.30% | Percentile11.11% | Participants28 | EvidenceC | Evaluated |
| Rank26 | ModelAC | Score45.30% | Percentile7.41% | Participants28 | EvidenceC | Evaluated |
| Rank27 | ModelAM Nova ProAmazon | Score41.60% | Percentile3.70% | Participants28 | EvidenceC | Evaluated |
| Rank28 | ModelAMNova LiteAmazon | Score40.40% | Percentile0.00% | Participants28 | EvidenceC | Evaluated |
LVBench Highlights
The leading models and scores on this benchmark.
LVBench Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for LVBench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is LVBench?
What LVBench measures and how its scores work.
LVBench is an extreme long video understanding benchmark for multimodal models, with videos averaging five times longer than existing datasets.
It measures video understanding on videos up to two hours in duration using the Score metric, reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- LVBench
- 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 LVBench.
Which model scores highest on LVBench?
Gemini 3.8 Flash is currently ranked first with 87.10%.
What are the top three models on LVBench?
The current leaders are Gemini 3.8 Flash (87.10%), Gemini 3.7 Flash (85.40%), and Qwen3.8 Max (81.80%).
Which LVBench model has the lowest official input price?
Nova Lite has the lowest matched official input price at $0.06 input / $0.24 output per 1M tokens.
Which models are fastest among LVBench results?
The fastest matched records are Gemini 3.7 Flash (117.86 tok/s via Google), Qwen3.8 Max (60.87 tok/s via DeepInfra), and Gemini 3.8 Flash (51.56 tok/s via Google).
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.