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
1DeepSeek V4 Pro 0813low· Context Arena data100%
2Gemini 3.1 Pro Previewhigh· Context Arena data100%
3Gemini 3.7 Flashhigh· Context Arena data100%
4Gemini 3.7 Flashmedium· Context Arena data100%
5GPT 5.4high· Context Arena data100%
6GPT 5.4xhigh· Context Arena data100%
7GPT 5.5high· Context Arena data100%
8GPT 5.5low· Context Arena data100%
9GPT 5.5medium· Context Arena data100%
10GPT 5.5xhigh· Context Arena data100%
11GPT-5.6 Solmax· Context Arena data100%
12Qwen 3.8 27Bxhigh· Context Arena data100%
13Qwen3.5 27Benabled· Context Arena data100%
14Qwen3.6 Max Previewenabled· Context Arena data100%
15Ling 3.0 Flashenabled· Context Arena data99.8%
16GLM 5.3max· Context Arena data99.7%
17Qwen 3.8 27Blow· Context Arena data99.3%
18Claude Opus 4.6low· Context Arena data99.3%
19Claude Opus 4.6medium· Context Arena data99.3%
20Qwen 3.8 27Bmedium· Context Arena data99.3%
21Qwen3.8 Maxlow· Context Arena data99.3%
22Qwen3.8 Maxmedium· Context Arena data99.3%
23Qwen3.6 35B A3Benabled· Context Arena data98.8%
24Gemini 3.1 Flash Litehigh· Context Arena data98.8%
25DeepSeek V4 Pro 0813high· Context Arena data98.8%
26Hy3 (Deep Infra)low· Context Arena data98.8%
27Gemini 3.1 Pro Previewlow· Context Arena data98.8%
28Gemini 3.6 Flashhigh· Context Arena data98.8%
29GPT-5.6 Terramax· Context Arena data98.8%
30Qwen3.6 Plusenabled· Context Arena data98.8%
31Claude Opus 5max· Context Arena data98.7%
32Qwen3.8 Maxxhigh· Context Arena data98.7%
33Grok 4.6medium· Context Arena data98.3%
34Claude Opus 4.6high· Context Arena data98.2%
35Grok 4.6low· Context Arena data97.6%
36Muse Spark 1.2high· Context Arena data97.6%
37Grok 4.6xhigh· Context Arena data97.6%
38Qwen3.6 27Benabled· Context Arena data97.6%
39Doubao Seed 2.1 Turboenabled· Context Arena data97.6%
40Gemini 3 Flashhigh· Context Arena data97.6%
41Gemini 3.7 Flashlow· Context Arena data97.6%
42Qwen3.7 Flashenabled· Context Arena data97.6%
43Muse Spark 1.2low· Context Arena data97.5%
44Gemini 3 Flashmedium· Context Arena data97.5%
45Kimi K3max· Context Arena data97.5%
46Claude Opus 4.8max· Context Arena data97.4%
47Qwen3.5 Plusenabled· Context Arena data96.9%
48Claude Sonnet 4.6high· Context Arena data96.7%
49GPT 5.4medium· Context Arena data96.5%
50GPT-5.6 Lunamax· Context Arena data96.4%

Showing 1–50 of 177 results.