Skip to content

Benchmark

GDM-MRCRv2: 8 needle 16k

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
2.62

Score distribution

69.7%
Mean
79.3%
Median
25.6%
Std dev
34.2%
10th percentile
97.5%
90th percentile
13.6%
Min
13.6%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.62 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
1Gemini 3.7 Flashmedium· Context Arena data100%
2GPT-5.6 Solmax· Context Arena data100%
3DeepSeek V4 Pro 0813high· Context Arena data100%
4Gemini 3.7 Flashhigh· Context Arena data100%
5GPT-5.6 Terramax· Context Arena data100%
6Claude Opus 5max· Context Arena data99.9%
7GPT 5.5xhigh· Context Arena data99.9%
8Hy3 (Deep Infra)high· Context Arena data99.9%
9Qwen 3.8 27Bmedium· Context Arena data99.2%
10Qwen 3.8 27Bxhigh· Context Arena data99%
11Gemini 3.6 Flashmedium· Context Arena data98.8%
12GPT 5.5high· Context Arena data98.8%
13GPT 5.4xhigh· Context Arena data98.8%
14Gemini 3.6 Flashhigh· Context Arena data98.8%
15Gemini 3.1 Pro Previewhigh· Context Arena data98.8%
16Qwen3.8 Maxmedium· Context Arena data98%
17DeepSeek V4 Pro 0813low· Context Arena data97.6%
18Hy3 (Deep Infra)low· Context Arena data97.6%
19Gemini 3.5 Flashhigh· Context Arena data97.5%
20Qwen3.8 Maxxhigh· Context Arena data97.5%
21Grok 4.6medium· Context Arena data96.9%
22Qwen3.5 397B A17Benabled· Context Arena data96.7%
23Qwen3.8 Maxlow· Context Arena data96.7%
24Muse Spark 1.2minimal· Context Arena data96.4%
25GPT 5.5low· Context Arena data96.3%
26GPT 5.5medium· Context Arena data96.3%
27Claude Opus 4.8max· Context Arena data96.1%
28DeepSeek V4 Proenabled· Context Arena data96.1%
29Grok 4.6xhigh· Context Arena data96.1%
30Gemini 3 Flashmedium· Context Arena data95.7%
31Qwen 3.8 27Blow· Context Arena data95.6%
32Gemini 3.1 Pro Previewlow· Context Arena data95.2%
33Muse Spark 1.2medium· Context Arena data95.2%
34Muse Spark 1.2low· Context Arena data95.1%
35Qwen3.6 35B A3Benabled· Context Arena data95.1%
36Gemini 3.5 Flashmedium· Context Arena data95.1%
37Grok 4.6low· Context Arena data95.1%
38Claude Sonnet 5max· Context Arena data94.9%
39GLM 5.3high· Context Arena data94.9%
40Qwen3.5 Plusenabled· Context Arena data94.3%
41Qwen3.5 27Benabled· Context Arena data94%
42Qwen3.6 27Benabled· Context Arena data94%
43Gemini 3.6 Flashlow· Context Arena data93.9%
44GPT-5.6 Lunamax· Context Arena data93.9%
45Qwen3.5 122B-A10Benabled· Context Arena data93.7%
46Qwen3.6 Plusenabled· Context Arena data92.8%
47Gemini 3.7 Flashlow· Context Arena data92.7%
48Muse Spark 1.2xhigh· Context Arena data92.7%
49Qwen3.5 35B A3Benabled· Context Arena data92.7%
50GLM 5.3max· Context Arena data92.7%

Showing 1–50 of 177 results.