# LongMemEval

> Tests whether a chat assistant recalls and reasons over information spread across long, multi-session conversation histories.

Canonical page: https://benchlm.ai/benchmarks/longmemeval

- Category: [Reasoning](/reasoning)
- Last updated: October 29, 2026

## About LongMemEval

- Year: 2024
- Tasks: Questions over multi-session chat histories
- Format: Overall accuracy
- Difficulty: Recall across conversations
- Paper: [LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory](https://github.com/xiaowu0162/LongMemEval)

NVIDIA reports it for Nemotron 3.5 Bolt E4B (October 23, 2026), run in its NeMo Gym harness, as overall accuracy in reasoning mode.

LongMemEval is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (2 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [Nemotron 3.5 Bolt E4B BF16](/models/nemotron-3-5-bolt-e4b-bf16) | NVIDIA | 72.10% |
| 2 | [Nemotron 3.5 Bolt E4B](/models/nemotron-3-5-bolt-e4b) | NVIDIA | 69.92% |

## FAQ

### What does LongMemEval measure?

Tests whether a chat assistant recalls and reasons over information spread across long, multi-session conversation histories.

### Which model scores highest on LongMemEval?

Nemotron 3.5 Bolt E4B BF16 by NVIDIA currently leads with a score of 72.10% on LongMemEval.

### How many models are evaluated on LongMemEval?

2 AI models have published results on LongMemEval in the BenchLM catalog.

### Does LongMemEval affect BenchLM's overall score?

Not directly. LongMemEval is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Compare Top Models on LongMemEval

- [Nemotron 3.5 Bolt E4B BF16 vs Nemotron 3.5 Bolt E4B](/compare/nemotron-3-5-bolt-e4b-vs-nemotron-3-5-bolt-e4b-bf16)