reasoning benchmark
LongCodeBench
LongCodeBench evaluates the code understanding and comprehension abilities of large language models at very long context windows, scaling up to 1M tokens. It tests whether models can reason about extensive codebases provided in a single prompt by answering multiple-choice questions about the code.
Updated Aug 7, 2026
Published models2
Registry coverage2
MetricScore
EvidenceB
LongCodeBench leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AM Nova 2 LiteAmazon | 84.0% | 100.0% | 2 | C | |
| 2 | AMNova 2 ProAmazon | 84.0% | 0.0% | 2 | C |
Score distribution
Top published rows on the benchmark's original scale.
LongCodeBench highlights
The top published results on this benchmark's own scale.
What is LongCodeBench?
Definition and scoring fields from the benchmark registry.
LongCodeBench evaluates the code understanding and comprehension abilities of large language models at very long context windows, scaling up to 1M tokens. It tests whether models can reason about extensive codebases provided in a single prompt by answering multiple-choice questions about the code.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- LongCodeBench
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- longcodebench|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about LongCodeBench.
Which model scores highest on LongCodeBench?
Nova 2 Lite is currently ranked first with 84.0%.
What does LongCodeBench measure?
LongCodeBench evaluates the code understanding and comprehension abilities of large language models at very long context windows, scaling up to 1M tokens. It tests whether models can reason about extensive codebases provided in a single prompt by answering multiple-choice questions about the code.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
2 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.