Skip to content

Benchmark

GDM-MRCRv2: 8 needle 64k

Long context benchmarks measure how well models use information in long inputs. Scores are percentages.

Models measured
68
Scores recorded
177
Sources
1
IQ weight
1.74

Score distribution

50.4%
Mean
49.6%
Median
24.4%
Std dev
20%
10th percentile
84.1%
90th percentile
8.9%
Min
8.9%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.74 for chat models, measured over 68 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51Ling 3.0 Flashenabled· Context Arena data67.5%
52Muse Spark 1.2low· Context Arena data66.9%
53Qwen3.5 397B A17Benabled· Context Arena data66.9%
54DeepSeek V4 Proenabled· Context Arena data66.5%
55Kimi K3max· Context Arena data66.5%
56Qwen3.6 Plusenabled· Context Arena data66%
57Qwen3.6 Max Preview· Context Arena data65.7%
58Muse Spark 1.2medium· Context Arena data64.7%
59DeepSeek V4 Flash 0731max· Context Arena data64.5%
60DeepSeek V4 Flashenabled· Context Arena data64.4%
61Ring 2.6 1Txhigh· Context Arena data63.9%
62Ring 2.6 1Thigh· Context Arena data63.6%
63Muse Spark 1.1medium· Context Arena data63.6%
64Muse Spark 1.1low· Context Arena data63.5%
65GLM 5.2max· Context Arena data63.4%
66Gemini 3.6 Flashlow· Context Arena data62.3%
67Qwen3.6 Plus· Context Arena data62.3%
68Gemini 3.5 Flashhigh· Context Arena data61.2%
69Qwen3.5 Plusenabled· Context Arena data60%
70Gemini 3.1 Pro Previewlow· Context Arena data60%
71GPT 5.4low· Context Arena data59.9%
72Gemini 3 Flashhigh· Context Arena data59.7%
73Kimi K3high· Context Arena data58.9%
74Claude Opus 4.6· Context Arena data58.4%
75Muse Spark 1.1minimal· Context Arena data57.9%
76Grok 4.6low· Context Arena data57.8%
77Gemini 3.5 Flash Litehigh· Context Arena data57.7%
78GPT 5.4medium· Context Arena data56.8%
79Gemini 3 Flashmedium· Context Arena data56.7%
80Qwen3.7 Max· Context Arena data56.6%
81Gemini 3.5 Flashmedium· Context Arena data55.3%
82Claude Sonnet 4.6· Context Arena data55%
83Qwen3.5 35B A3Benabled· Context Arena data52.9%
84Qwen3.5 122B-A10Benabled· Context Arena data51.9%
85Qwen3.6 27B· Context Arena data51.6%
86Gemini 3.5 Flash Litemedium· Context Arena data50.8%
87Nemotron 3 Super 120B A12Benabled· Context Arena data50.7%
88Qwen3.5 27Benabled· Context Arena data49.6%
89Muse Spark 1.1xhigh· Context Arena data49.6%
90Muse Spark 1.1high· Context Arena data49%
91Qwen3.6 Flash· Context Arena data48.6%
92Kimi K3low· Context Arena data48.6%
93Doubao Seed 2.1 Turboenabled· Context Arena data48.1%
94Qwen3.7 Flash· Context Arena data47.5%
95Kimi K2.6enabled· Context Arena data46.3%
96Gemini 3.5 Flash Litelow· Context Arena data45.2%
97GLM 5.1enabled· Context Arena data45%
98Kimi K2.6· Context Arena data43.5%
99Qwen 3.8 27B· Context Arena data42.9%
100Hy3 (Deep Infra)low· Context Arena data42.8%

Showing 51–100 of 177 results.