# AI-Needle leaderboard

Reasoning benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/leaderboards/benchmarks/reasoning-ai-needle

Every model the catalog carries a published AI-Needle value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Reasoning |
| Measure | Needle-in-a-haystack recall |
| Tasks | Long-context retrieval |
| Difficulty | Long-context memory |
| Published by | [Qwen3.6 launch benchmarks](https://qwen.ai/blog?id=qwen3.6) |

A long-context retrieval benchmark that measures whether a model can recover relevant information embedded deep inside very long contexts.

## AI-Needle ranking

4 models with a published AI-Needle value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Needle-in-a-haystack recall | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Claude Opus 4.5](https://modelscale.dev/models/claude-opus-4-5) | Anthropic | 74 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Qwen3.5 397B](https://modelscale.dev/models/qwen3-5-397b) | Alibaba | 68.7 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Qwen3.6 Plus](https://modelscale.dev/models/qwen3-6-plus) | Alibaba | 68.3 | Observed 2026-11-21 · source benchlm:models |
| 4 | [GLM-5](https://modelscale.dev/models/glm-5) | Z.AI | 63.3 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
