Skip to content

Benchmark

JMLE2026-Bench (120th JMLE): exam points, all questions

Science benchmarks measure how well models answer science questions. Higher scores are better.

Models measured
15
Scores recorded
27
Sources
1
IQ weight
1.97

Score distribution

411
Mean
449
Median
98.8
Std dev
243
10th percentile
486
90th percentile
156
Min
156493
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.97 for chat models, measured over 15 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 4.6· JMLE2026-Bench and IgakuQA119 data docs493
2Gemini 3.1 Pro Preview· JMLE2026-Bench and IgakuQA119 data docs493
3Claude Sonnet 4.6· JMLE2026-Bench and IgakuQA119 data docs489
4GPT 5.2· JMLE2026-Bench and IgakuQA119 data docs486
5Qwen3.5 35B A3B· JMLE2026-Bench and IgakuQA119 data docs480
6Qwen3.5 397B A17B· JMLE2026-Bench and IgakuQA119 data docs480
7Qwen3.5 122B-A10B· JMLE2026-Bench and IgakuQA119 data docs479
8Qwen3.5 27B· JMLE2026-Bench and IgakuQA119 data docs475
9GPT OSS 120Bhigh· JMLE2026-Bench and IgakuQA119 data docs468
10Qwen3.5 27Bno-think· JMLE2026-Bench and IgakuQA119 data docs463
11Qwen3.5 122B-A10Bno-think· JMLE2026-Bench and IgakuQA119 data docs458
12Qwen3.5 35B A3Bno-think· JMLE2026-Bench and IgakuQA119 data docs455
13GPT OSS 120Bmedium· JMLE2026-Bench and IgakuQA119 data docs453
14Qwen3 32B· JMLE2026-Bench and IgakuQA119 data docs449
15Qwen3.5 9B· JMLE2026-Bench and IgakuQA119 data docs436
16GPT OSS 120Blow· JMLE2026-Bench and IgakuQA119 data docs434
17GPT OSS 20bmedium· JMLE2026-Bench and IgakuQA119 data docs429
18GPT OSS 20bhigh· JMLE2026-Bench and IgakuQA119 data docs425
19Qwen3 32Bno-think· JMLE2026-Bench and IgakuQA119 data docs417
20Qwen3.5 9Bno-think· JMLE2026-Bench and IgakuQA119 data docs415
21Qwen3.5 4B· JMLE2026-Bench and IgakuQA119 data docs409
22GPT OSS 20blow· JMLE2026-Bench and IgakuQA119 data docs386
23Qwen3.5 4Bno-think· JMLE2026-Bench and IgakuQA119 data docs370
24Qwen3.5 2Bno-think· JMLE2026-Bench and IgakuQA119 data docs243
25Qwen3.5 2B· JMLE2026-Bench and IgakuQA119 data docs195
26Qwen3.5 0.8Bno-think· JMLE2026-Bench and IgakuQA119 data docs160
27Qwen3.5 0.8B· JMLE2026-Bench and IgakuQA119 data docs156

Data credits: JMLE2026-Bench and IgakuQA119 by Naoto Iwase; results licence not stated (code MIT, question data CC BY 4.0).