Skip to content

Benchmark

GDM-MRCRv2: 8 needle 8k

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
2.96

Score distribution

79.6%
Mean
89.2%
Median
21.3%
Std dev
47.2%
10th percentile
99.3%
90th percentile
19.8%
Min
19.8%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.96 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
51Claude Sonnet 5max· Context Arena data96.4%
52Hy3 (Deep Infra)high· Context Arena data96.4%
53GPT 5.4low· Context Arena data96.4%
54DeepSeek V4 Proenabled· Context Arena data96.3%
55DeepSeek V4 Flashenabled· Context Arena data96.3%
56Gemini 3.1 Flash Litemedium· Context Arena data96.3%
57Gemini 3.5 Flash Litemedium· Context Arena data96.3%
58DeepSeek V4 Flash 0731max· Context Arena data96.3%
59Gemini 3.5 Flash Litehigh· Context Arena data96.3%
60GLM 5.1enabled· Context Arena data96%
61GLM 5.3 Flashmax· Context Arena data96%
62GLM 5.3high· Context Arena data96%
63Grok 4.6high· Context Arena data95.9%
64Claude Sonnet 4.6medium· Context Arena data95.8%
65Grok 4.20enabled· Context Arena data95.2%
66DeepSeek V4 Flash 0731high· Context Arena data95.2%
67Kimi K2.6enabled· Context Arena data95.1%
68Muse Spark 1.2minimal· Context Arena data95.1%
69Kimi K2.5enabled· Context Arena data95.1%
70GLM 5.2max· Context Arena data95.1%
71Qwen3.5 122B-A10Benabled· Context Arena data95.1%
72Gemini 3.5 Flashhigh· Context Arena data95.1%
73Claude Sonnet 4.6low· Context Arena data94.8%
74Qwen3.5 397B A17Benabled· Context Arena data94.4%
75Gemini 3.6 Flashmedium· Context Arena data93.9%
76DeepSeek V4 Pro 0813max· Context Arena data93.9%
77GLM 5.3 Flashhigh· Context Arena data93.8%
78Qwen3.5 35B A3Benabled· Context Arena data93.8%
79Muse Spark 1.2xhigh· Context Arena data92.7%
80Muse Spark 1.2medium· Context Arena data92.7%
81Kimi K3low· Context Arena data91.6%
82Kimi K3high· Context Arena data91.5%
83MiniMax M3enabled· Context Arena data91.3%
84Ring 2.6 1Thigh· Context Arena data91%
85Gemini 3.5 Flashmedium· Context Arena data90.4%
86Muse Spark 1.1minimal· Context Arena data90.3%
87Qwen3.5 9Benabled· Context Arena data90.2%
88GLM 5.3low· Context Arena data90%
89Muse Spark 1.1high· Context Arena data89.2%
90Muse Spark 1.1medium· Context Arena data89.2%
91Claude Opus 4.7xhigh· Context Arena data89%
92Gemini 3.5 Flashlow· Context Arena data87.9%
93Muse Spark 1.1low· Context Arena data87.8%
94Gemini 3.5 Flash Litelow· Context Arena data87.7%
95Claude Opus 4.7low· Context Arena data86.5%
96Muse Spark 1.1xhigh· Context Arena data86.5%
97DeepSeek V4 Flash 0731low· Context Arena data86.2%
98GPT 5.4 Nanoxhigh· Context Arena data85.7%
99Gemini 3 Flashlow· Context Arena data84.5%
100Qwen3.6 Max Preview· Context Arena data84.4%

Showing 51–100 of 177 results.