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
101MiMo V2.5 Proenabled· Context Arena data82.9%
102Claude Haiku 4.5enabled· Context Arena data82.5%
103Claude Opus 4.7medium· Context Arena data82%
104GLM 5.3 Flashlow· Context Arena data81.7%
105Claude Opus 4.6· Context Arena data81.6%
106Ring 2.6 1Txhigh· Context Arena data81.6%
107Gemini 3.6 Flashlow· Context Arena data80.5%
108Qwen3.7 Flash· Context Arena data79.5%
109MiMo V2.5enabled· Context Arena data79.4%
110Claude Opus 4.7high· Context Arena data78.8%
111GPT-5.4 minixhigh· Context Arena data77.1%
112Qwen3.6 Plus· Context Arena data76%
113Kimi K2.5· Context Arena data75.9%
114GPT 5.4 Nanohigh· Context Arena data74.5%
115Claude Sonnet 4.6· Context Arena data74.3%
116MiMo V2 Proenabled· Context Arena data72.2%
117Inkling Smallmax· Context Arena data72.2%
118Mistral Medium 3.5high· Context Arena data71.7%
119Qwen3.6 35B A3B· Context Arena data71.1%
120Kimi K2.6· Context Arena data70.7%
121MiMo V2 Omnienabled· Context Arena data70.6%
122Step 3.7 Flashlow· Context Arena data69.3%
123GPT-5.4 minihigh· Context Arena data68.8%
124Qwen3.6 27B· Context Arena data68.6%
125Qwen3.6 Flash· Context Arena data68.5%
126Inkling Smallhigh· Context Arena data68.4%
127GPT-5.4 minimedium· Context Arena data67.7%
128GPT 5.5· Context Arena data66.4%
129Ling 3.0 Flash· Context Arena data66.1%
130Qwen3.7 Max· Context Arena data66.1%
131Step 3.7 Flashmedium· Context Arena data65.4%
132Gemini 3.1 Flash Litelow· Context Arena data65.2%
133Doubao Seed 2.1 Turbo· Context Arena data63.9%
134GPT-5.4 minilow· Context Arena data63.6%
135MiniMax M2.7enabled· Context Arena data63.6%
136DeepSeek V4 Pro 0813· Context Arena data62.6%
137GPT 5.4 Nanomedium· Context Arena data61.4%
138Step 3.7 Flashhigh· Context Arena data61.3%
139Qwen3.5 Plus· Context Arena data61.1%
140Qwen 3.8 27B· Context Arena data60.3%
141Gemini 3.5 Flash· Context Arena data60.1%
142Qwen3.5 27B· Context Arena data59%
143Mistral Smallenabled· Context Arena data58.7%
144Gemini 3 Flash· Context Arena data58.7%
145GPT 5.4· Context Arena data57.8%
146Inkling Smallmedium· Context Arena data56.5%
147Qwen3.5 122B-A10B· Context Arena data55.3%
148Qwen3.5 397B A17B· Context Arena data55.1%
149Inkling Smalllow· Context Arena data54.3%
150Nemotron 3 Super 120B A12Benabled· Context Arena data54.1%

Showing 101–150 of 177 results.