Support

Settings

Theme

Explore themes

Mode

Library

Looks borrowed from benchmark sites and products. Themes with a light and a dark version follow your mode.

Bring your own theme

Upload a Priors theme file, or have your agent make one: copy the prompt, tell it the look you want, and upload what it returns. With the Priors skill installed it already knows the format.

Custom themes stay in this browser. They travel inside shared chart links and a page's theme, so others see them as you do.

For the builder, Explore and model pages on this device. Published pages keep the theme their owner picked.

Mode

Update notices

A note when a benchmark one of your indexes uses gets a new version, with a one-click upgrade.

Benchmarks · Reasoning & puzzles · Knowledge & accuracy

LisanBench

Word chains: from each of 50 starting words, the model links real words that each differ from the last by one letter (added, removed or swapped), never repeating a word, for as long as it can. Tests vocabulary, planning and rule-following. Valid steps per chain, summed over the 50 words (three tries each). By @scaling01.

116 models · 190 results · data as of 2026-11-16 · lisanbench.com ↗

Leaderboard

Muse Wonder Full xhigh

More from this benchmark

4 published measures

Breakdown

darker is better within each column · click a column to sort
Model
Muse Wonder Full xhigh9,3371149.0
Muse Wonder max8,3030.999150.0
Opus 5.5 xhigh5,9630.999150.0
DeepSeek V4.1 Flash 1017 max7,7330.988150.0
Opus 5 high5,0680.998150.0
Fable 5.1 xhigh4,5730.995150.0
Opus 4.7 xhigh5,1230.997149.0
Opus 4.63,5260.988150.0
Sonnet 5.5 xhigh2,9830.999150.0
Fable 5 medium4,5620.998150.0
Haiku 5.5 xhigh2,7110.997150.0
Sonnet 4.62,9440.993150.0
Kimi K3 max3,5260.997150.0
GPT-5.5 medium3,3160.995150.0
Opus 4.8 high2,6940.999150.0
GPT-6.1 Sol high4,0040.98150.0
GPT-6 Sol high3,1180.988150.0
Gemini 4 Argon high3,8490.753150.0
GPT-6 Astra medium3,1130.989150.0
Gemini 3.1 Pro high1,9290.96150.0
Sonnet 5 high1,7370.999150.0
Opus 4.52,2040.947150.0
GPT-5.4 medium2,7380.962150.0
GPT-5.6 Sol medium1,4030.979150.0
Gemini 3 Pro high1,1310.955150.0

All results

190 rows
#ModelEffortLisanBench
1Muse Wonder Fullxhigh32190.2 ±1358.2
2Muse Wondermax27973.3 ±1312.4
3Muse Wonder Fullhigh24760.3 ±1207.3
4Opus 5.5xhigh24029.0 ±1316.5
5Muse Wonderxhigh22588.7 ±1300.2
6DeepSeek V4.1 Flash 1017max21107.3 ±1290.0
7Muse Wonder Fullmedium18245.3 ±795.3
8Muse Wonderhigh18145.3 ±979.0
9Opus 5high15428.3 ±1044.4
10Fable 5.1xhigh15100.3 ±1557.4
11DeepSeek V4.1 Flash 1017high14750.0 ±968.4
12Opus 4.7xhigh14400.0 ±749.4
13Opus 4.614206.3 ±678.8
14Muse Wondermedium13807.0 ±754.9
15Sonnet 5.5xhigh13787.0 ±712.4
16Fable 5medium13680.3 ±1018.5
17Muse Wonder Fulllow13338.3 ±794.9
18Haiku 5.5xhigh12897.0 ±1019.4
19Sonnet 4.612060.0 ±563.0
20DeepSeek V4.1 Flash 1017low11638.3 ±866.2
21Kimi K3max10520.7 ±768.9
22GPT-5.5medium9844.7 ±854.5
23Opus 4.8high9463.0 ±766.3
24Muse Wonderlow8833.3 ±620.4
25GPT-6.1 Solhigh8736.0 ±1557.0
26Opus 5.5high8660.3 ±446.7
27GPT-6 Solhigh8222.0 ±1068.0
28Fable 5.1high8171.3 ±628.7
29Opus 5medium8112.7 ±651.7
30Haiku 5.5high7608.0 ±563.8