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

Benchmarks · math

MGSM (Vals)

A multilingual benchmark for mathematical questions. Scores are the values published by VALS AI. Source benchmark version 1.

11 models · 11 results · updated 2026-01-09 · vals.ai ↗

Leaderboard

Chart settings
11 of 11 selected
DeepSeek R1 xhigh

Score vs subscription cost per task

MGSM (Vals) · measured cost per task · subscription-priced where plan data exists (Real API Pricing, full allowance used)

Highlight models

Click a point on the chart to highlight that exact effort; click it again to remove it.

3 more in settings
Chart settings
11 of 11 selected

More from this benchmark

12 published measures

Breakdown

darker is better within each column · click a column to sort
Model
DeepSeek R1 xhigh98.091.691.688.493.292.090.894.494.889.290.8
DeepSeek V3 none97.690.891.691.693.693.291.692.096.087.288.4
gpt-oss-120b high98.088.892.492.895.292.883.296.495.288.888.8
Qwen 3 Max98.490.492.089.695.693.286.094.493.287.689.6
Grok 3 none97.286.492.482.895.692.090.893.294.888.091.6
GLM 4.5 xhigh96.090.090.889.694.492.884.892.892.486.489.2
GLM 4.6 none99.289.691.274.494.892.082.892.491.288.491.2
gpt-oss-20b high97.689.290.884.095.293.284.490.491.686.476.4
GLM 4.7 none98.489.692.456.896.893.282.093.691.286.489.6
Grok 4 Fast none95.284.889.278.492.891.283.691.692.085.284.0
MiniMax M2.1100.090.090.458.893.293.683.692.492.084.488.0

All results

11 rows
#ModelEffortMGSM (Vals)Subscription $ / taskAPI $ / taskTime
1DeepSeek R1xhigh92.3%$0.004$0.00410s
2DeepSeek V3none92.1%$0.000$0.00016s
3gpt-oss-120bhigh92.0%$0.001$0.0015.8s
4Qwen 3 Maxmax91.8%$0.003$0.00311s
5Grok 3none91.3%$0.006$0.0065.8s
6GLM 4.5xhigh90.8%$0.006$0.00690s
7GLM 4.6none89.7%$0.004$0.00434s
8gpt-oss-20bhigh89.0%$0.000$0.0005.8s
9GLM 4.7none88.2%$0.004$0.00463s
10Grok 4 Fastnone88.0%$0.000$0.0001.8s
11MiniMax M2.187.9%$0.001$0.00114s