# MRCR v2 (8-needle) (1M pointwise) — AI model rankings

Tests whether the AI can find specific details buried inside an enormous document (around 1 million tokens — many books). Higher is better.

3 tracked models have a published MRCR v2 (8-needle) (1M pointwise) score. Higher is better. Scores are as published at each model's release.

## Ranking

| Rank | Model | Developer | Score | Released |
| --- | --- | --- | --- | --- |
| 1 | Gemini 3.5 Flash | Google | 26.6% | May 19 2026 |
| 2 | Gemini 3.1 Pro | Google | 26.3% | Feb 19 2026 |
| 3 | Gemini 3.0 Flash | Google | 22.1% | Dec 17 2025 |


---

Canonical page: https://aireleasetracker.com/benchmark/mrcr-v2-1m
Full dataset: https://aireleasetracker.com/llms-full.txt · JSON: https://aireleasetracker.com/models.json
Source: AI Release Tracker (https://aireleasetracker.com). Benchmark scores are the figures published by the releasing lab at launch.