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
1Qwen3.8 Maxxhigh· Context Arena data99.3%
2Claude Opus 5max· Context Arena data99.3%
3Gemini 3.7 Flashhigh· Context Arena data98.6%
4Gemini 3.7 Flashmedium· Context Arena data98.5%
5Qwen3.8 Maxmedium· Context Arena data98.1%
6GPT-5.6 Solmax· Context Arena data98%
7DeepSeek V4 Pro 0813max· Context Arena data97.2%
8Gemini 3.7 Flashlow· Context Arena data97.2%
9Qwen 3.8 27Bmedium· Context Arena data96.7%
10GPT 5.5high· Context Arena data96.6%
11GPT 5.5xhigh· Context Arena data95.8%
12Qwen3.6 Max Previewenabled· Context Arena data95.7%
13Qwen 3.8 27Blow· Context Arena data95.3%
14Qwen3.8 Maxlow· Context Arena data94.8%
15DeepSeek V4 Pro 0813high· Context Arena data94.5%
16GLM 5.3high· Context Arena data94.3%
17GPT-5.6 Terramax· Context Arena data94.3%
18Qwen3.6 35B A3Benabled· Context Arena data94.3%
19Gemini 3.1 Pro Previewhigh· Context Arena data94.2%
20Claude Opus 4.8max· Context Arena data93%
21Qwen 3.8 27Bxhigh· Context Arena data92.9%
22Qwen3.6 27Benabled· Context Arena data92.8%
23Muse Spark 1.2xhigh· Context Arena data91.8%
24GPT 5.5low· Context Arena data91.8%
25GPT 5.5medium· Context Arena data91.6%
26Muse Spark 1.2high· Context Arena data91.5%
27GPT 5.4xhigh· Context Arena data91.5%
28Grok 4.6medium· Context Arena data91.5%
29GLM 5.3max· Context Arena data91.2%
30Muse Spark 1.2medium· Context Arena data90.9%
31Gemini 3.6 Flashhigh· Context Arena data90.8%
32GPT-5.6 Lunamax· Context Arena data90.2%
33Gemini 3.6 Flashmedium· Context Arena data90.1%
34Grok 4.6xhigh· Context Arena data89.6%
35Muse Spark 1.2low· Context Arena data89.5%
36DeepSeek V4 Pro 0813low· Context Arena data89.4%
37Grok 4.6high· Context Arena data89.1%
38Claude Opus 4.6low· Context Arena data88.2%
39Claude Sonnet 5max· Context Arena data87.9%
40GPT 5.4high· Context Arena data87.5%
41Claude Opus 4.6medium· Context Arena data87.5%
42Claude Opus 4.6high· Context Arena data86.2%
43GLM 5.3 Flashhigh· Context Arena data85.9%
44Claude Sonnet 4.6low· Context Arena data85.8%
45Muse Spark 1.1xhigh· Context Arena data85.8%
46Muse Spark 1.1medium· Context Arena data85.2%
47Claude Sonnet 4.6high· Context Arena data85.2%
48Qwen3.7 Flashenabled· Context Arena data85.1%
49Qwen3.6 Plusenabled· Context Arena data84.5%
50Gemini 3.1 Pro Previewlow· Context Arena data84.4%

Showing 1–50 of 177 results.