Skip to content

Benchmark

GDM-MRCRv2: 8 needle 32k

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
1.85

Score distribution

60.7%
Mean
65.5%
Median
27.2%
Std dev
23.8%
10th percentile
94.2%
90th percentile
11.8%
Min
11.8%99.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.85 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
51GPT 5.4medium· Context Arena data84%
52Claude Sonnet 4.6medium· Context Arena data83.8%
53Qwen3.5 27Benabled· Context Arena data82.9%
54Ling 3.0 Flashenabled· Context Arena data82.8%
55Qwen3.5 397B A17Benabled· Context Arena data82.8%
56Grok 4.6low· Context Arena data82.5%
57GLM 5.3 Flashmax· Context Arena data82.1%
58DeepSeek V4 Flash 0731max· Context Arena data80.3%
59Qwen3.5 Plusenabled· Context Arena data79.7%
60Muse Spark 1.1minimal· Context Arena data79.6%
61Muse Spark 1.2minimal· Context Arena data79.6%
62Gemini 3.5 Flashhigh· Context Arena data79.6%
63DeepSeek V4 Proenabled· Context Arena data79.4%
64Muse Spark 1.1high· Context Arena data78.2%
65Gemini 3.5 Flashmedium· Context Arena data78.1%
66GLM 5.3low· Context Arena data77.7%
67GPT 5.4low· Context Arena data76.8%
68DeepSeek V4 Flash 0731high· Context Arena data76.6%
69Gemini 3.5 Flash Litehigh· Context Arena data76.1%
70Gemini 3 Flashhigh· Context Arena data75.2%
71GLM 5.2max· Context Arena data74.7%
72Muse Spark 1.1low· Context Arena data74.6%
73Kimi K3max· Context Arena data74.5%
74Ring 2.6 1Txhigh· Context Arena data74.5%
75Qwen3.5 122B-A10Benabled· Context Arena data74%
76Hy3 (Deep Infra)low· Context Arena data73.9%
77Qwen3.5 35B A3Benabled· Context Arena data73.2%
78Gemini 3 Flashmedium· Context Arena data73%
79Hy3 (Deep Infra)high· Context Arena data72.9%
80GLM 5.3 Flashlow· Context Arena data72.7%
81Gemini 3.6 Flashlow· Context Arena data70.2%
82Qwen3.6 Max Preview· Context Arena data69.8%
83Ring 2.6 1Thigh· Context Arena data69.7%
84Doubao Seed 2.1 Turboenabled· Context Arena data69%
85GLM 5.1enabled· Context Arena data68.9%
86Kimi K2.6enabled· Context Arena data68.3%
87DeepSeek V4 Flashenabled· Context Arena data67.1%
88Gemini 3.1 Flash Litehigh· Context Arena data65.9%
89Kimi K3high· Context Arena data65.5%
90Gemini 3.5 Flash Litemedium· Context Arena data64.6%
91Gemini 3.5 Flash Litelow· Context Arena data63.4%
92Kimi K2.5enabled· Context Arena data62.7%
93Kimi K2.5· Context Arena data62%
94Grok 4.20enabled· Context Arena data61.9%
95Claude Opus 4.6· Context Arena data61.1%
96Qwen3.7 Max· Context Arena data60.5%
97Nemotron 3 Super 120B A12Benabled· Context Arena data60.3%
98Gemini 3.5 Flashlow· Context Arena data58.4%
99DeepSeek V4 Flash 0731low· Context Arena data58.1%
100Kimi K3low· Context Arena data57.1%

Showing 51–100 of 177 results.