long context benchmark
LongBench v2 Leaderboard
LongBench v2 is a benchmark for assessing LLMs on long-context problems requiring deep understanding and reasoning across real-world multitasks.
Updated Sep 5, 2026
Models17
Model coverage17
MetricScore
EvidenceC
LongBench v2 Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAC | Score66.30% | Percentile100.00% | Participants17 | EvidenceC | Evaluated |
| Rank02 | ModelAC | Score63.20% | Percentile93.75% | Participants17 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score62.00% | Percentile87.50% | Participants17 | EvidenceC | Evaluated |
| Rank04 | ModelNV | Score61.90% | Percentile81.25% | Participants17 | EvidenceC | Evaluated |
| Rank05 | ModelMI | Score61.50% | Percentile75.00% | Participants17 | EvidenceC | Evaluated |
| Rank06 | ModelMA | Score61.00% | Percentile68.75% | Participants17 | EvidenceC | Evaluated |
| Rank07 | ModelMI | Score61.00% | Percentile62.50% | Participants17 | EvidenceC | Evaluated |
| Rank08 | ModelMI | Score61.00% | Percentile56.25% | Participants17 | EvidenceC | Evaluated |
| Rank09 | ModelXI | Score60.60% | Percentile50.00% | Participants17 | EvidenceC | Evaluated |
| Rank10 | ModelAC | Score60.60% | Percentile43.75% | Participants17 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Score60.20% | Percentile37.50% | Participants17 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score59.00% | Percentile31.25% | Participants17 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Score55.20% | Percentile25.00% | Participants17 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Score50.00% | Percentile18.75% | Participants17 | EvidenceC | Evaluated |
| Rank15 | ModelDE | Score48.70% | Percentile12.50% | Participants17 | EvidenceC | Evaluated |
| Rank16 | ModelAC | Score38.70% | Percentile6.25% | Participants17 | EvidenceC | Evaluated |
| Rank17 | ModelAC | Score26.10% | Percentile0.00% | Participants17 | EvidenceC | Evaluated |
LongBench v2 Highlights
The leading models and scores on this benchmark.
LongBench v2 Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for LongBench v2
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is LongBench v2?
What LongBench v2 measures and how its scores work.
It consists of 503 challenging multiple-choice questions with contexts ranging from 8k to 2M words across six task categories: single-document QA, multi-document QA, long in-context learning, long-dialogue history understanding, code repository understanding, and long structured data understanding.
It measures performance on these six long-context task categories, reported as Score with unit ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of C.
- Family
- LongBench v2
- Modality
- text
- 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 LongBench v2.
Which model scores highest on LongBench v2?
Qwen3.8 Max is currently ranked first with 66.30%.
What are the top three models on LongBench v2?
The current leaders are Qwen3.8 Max (66.30%), Qwen3.5-397B-A17B (63.20%), and Qwen3.6 Plus (62.00%).
Which LongBench v2 model has the lowest official input price?
MiMo-V2-Flash has the lowest matched official input price at $0.14 input / $0.28 output per 1M tokens.
Which models are fastest among LongBench v2 results?
The fastest matched records are DeepSeek-V3 (100.00 tok/s via DeepSeek) and Qwen3.8 Max (60.87 tok/s via DeepInfra).
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.