BABILong (NIAH, v0, 16k): leaderboard
Metric: QA1–5 Mean Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 25 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.1 70B Instruct | 65 |
| 2 | GPT-4 Preview (0125) | 64 |
| 3 | Llama 3.1 8B Instruct | 60 |
| 4 | Phi-3-medium-128k-instruct | 57 |
| 5 | c4ai-command-r-v01 | 52 |
| 6 | Mixtral 8x22B Instruct (v0.1) | 51 |
| 7 | Yi-34B-200K | 50 |
| 8 | Mixtral 8x7B Instruct (v0.1) | 46 |
| 9 | Phi-3 Mini 128K Instruct | 46 |
| 10 | Mistral 7B Instruct (v0.2) | 42 |
| 11 | LongChat-7B-v1.5-32k | 39 |
| 12 | Yi-9B-200K | 36 |
| 13 | Llama 2 7B 32K Instruct | 35 |
| 14 | Llama 2 7B 32K | 32 |
| 15 | Yi 34B (Base) | 31 |
Interactive version: theaggregate.ai/benchmark?slug=babilong-niah-v0-16k · How It Works · Data refreshed daily, snapshot 2026-10-09.