Skip to content

Benchmark

Arena ELO Text Factuality - Math

Math benchmarks measure how well models solve maths problems. Higher scores are better.

Models measured
107
Scores recorded
122
Sources
1
IQ weight
2.16

Score distribution

1,450.8
Mean
1,446
Median
32.8
Std dev
1,411.7
10th percentile
1,493.9
90th percentile
1,358
Min
1,3581,518.4
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.16 for chat models, measured over 107 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 5Max· LMArena data1,518.4
2Claude Fable 5.1Max· LMArena data1,517.9
3Claude Opus 5High· LMArena data1,515.8
4Claude Opus 5.5High· LMArena data1,512.7
5Claude Fable 5High· LMArena data1,511.1
6Gemini 3.8 FlashHigh· LMArena data1,510
7Claude Opus 4.6High· LMArena data1,505.7
8Gemini 4 ArgonHigh· LMArena data1,505.3
9Claude Opus 4.6· LMArena data1,495.9
10Gemini 3.7 FlashHigh· LMArena data1,495.3
11Muse Spark 1.3Max· LMArena data1,494.8
12GLM 5.3 Flash· LMArena data1,494.8
13GPT 5.5· LMArena data1,493.9
14Claude Sonnet 5.5xHigh· LMArena data1,491.6
15Gemini 3.5 FlashHigh· LMArena data1,491.5
16Gemini 3.6 FlashHigh· LMArena data1,491.5
17GPT-6 AstraMax· LMArena data1,491.4
18GPT 5.4High· LMArena data1,490.4
19Claude Opus 4.7High· LMArena data1,490
20Qwen3.8 Max· LMArena data1,489.7
21GPT-5.6 SolxHigh· LMArena data1,486.9
22GPT 5.5High· LMArena data1,486
23Gemini 3.5 FlashMedium· LMArena data1,485
24Kimi K3Max· LMArena data1,483
25GPT-5.6 TerraxHigh· LMArena data1,479.5
26Claude Opus 4.8High· LMArena data1,479.3
27Claude Opus 4.7· LMArena data1,478.4
28Inkling FP4· LMArena data1,478.2
29MiMo V2.5 Pro· LMArena data1,477.5
30Gemini 3.1 Pro Preview· LMArena data1,476.4
31Gemini 3 Pro Preview· LMArena data1,475.9
32Grok 4.5· LMArena data1,474.3
33GLM 5.2Max· LMArena data1,474.3
34GPT 5.4· LMArena data1,474
35GLM 5.3Max· LMArena data1,473.1
36Kimi K2.6· LMArena data1,472.7
37ERNIE 5.1· LMArena data1,471
38Muse Spark 1.1· LMArena data1,467.9
39Claude Sonnet 5High· LMArena data1,467.8
40GPT-5.6 LunaxHigh· LMArena data1,467.6
41Claude Opus 4.8· LMArena data1,467.5
42Gemini 3 Flash· LMArena data1,467.1
43Hy3 (Deep Infra)· LMArena data1,464.6
44Qwen3.7 Plus· LMArena data1,462.4
45Claude Sonnet 4.6· LMArena data1,461.1
46Claude Opus 4.5High, 32K, 20251101· LMArena data1,460.6
47GLM 5.1· LMArena data1,460.1
48Claude Opus 4.520251101· LMArena data1,458
49Qwen3.5 397B A17B· LMArena data1,456.4
50MiMo V2.6 Flash· LMArena data1,455.6

Showing 1–50 of 122 results.

Data credits: LMArena, CC BY 4.0 (licence)