Towards AITowards AIToneBench

Lab Comparison

Mistral Models Compared

Last updated , from board data generated at 17:52 UTC.

All 4 ranked Mistral configurations on ToneBench, side by side: Writing Elo, Writing Score, the nine metrics, and cost and time per script. A configuration is one model at one thinking level.

Highest Writing Elo
Elo 1590Mistral Large 4 (high)#37 of 87 models · $0.089 per script
Lowest Cost
$0.0099 per scriptMistral Large 3 (no thinking)Configuration #135 · Elo 562

Recommended Configurations

Ranks in these notes are configuration ranks: each setting's place among all 151 configurations. The Highest Writing Elo card above gives the model rank.

Picks and Notes From Louis

If you're set on Mistral, use Mistral Large 4 (high). It costs $0.089 a script, and nothing else in the family comes close. Two things to know before you do. It's slow, at 8.7 minutes per script. And it still sits behind much cheaper models from other labs, so I'd only pick it if Mistral is a requirement.

I'd skip the rest for writing. Mistral Medium 3.5 (high) costs 2.3 times as much as Mistral Large 4 (high), going by its estimated price, and ranks #134. Turning its reasoning off makes it cheaper and no better. Mistral Large 3 (no thinking) is the cheapest at $0.0099, and at #135 that's a saving I wouldn't take.

Mistral has one script writer worth a look: Mistral Large 4 (high), at #77. The other Mistral rows are a long way down. The next one, Mistral Medium 3.5 (high), is at #134, and the family spans 1037 Elo from top to bottom. I like that Mistral ships open weights, and Mistral Large 3 (no thinking) is the open one here, but for scripts it lands at #135.

All 4 Mistral Configurations, Grouped by Model

Every row wrote the same 10 scripts for the same four blind judges, so these Writing Elo ratings sit on the leaderboard's scale. The rank column is the configuration rank, each setting's place among all 151 configurations on the board. Rows are grouped by model: a highlighted row is a model's best configuration, and the rows under it are the same model at other thinking levels. The ± after each Elo is half the width of its 95% range, and scores run from 0 to 100. Cost is one full script at the API list price of the route used. * marks an estimated price. Each name opens its scorecard, and the head-to-head comparisons cover models from different labs.

Score colors:82 and up65 to 81.9under 65

With the metric scores shown, the table scrolls sideways.

Configuration RankConfigurationWriting EloWriting ScoreCost / ScriptTime / ScriptTone & Voice MatchWriting Craft & ClaritySubstance, Accuracy & ValueHook StrengthYouTube Best PracticesContinuity & EmotionSlop ScoreLength AdherenceVisual Cue Quality
#77Mistral Large 4 (high)1590 ±5377.0$0.0898.7 min72.472.579.680.474.775.677.792.178.0
#134Mistral Medium 3.5 (high)577 ±6061.1$0.201*2.1 min55.255.767.173.763.558.262.560.356.2
#136Mistral Medium 3.5 (no thinking)553 ±4161.3$0.033*0.2 min55.763.366.475.757.961.371.440.564.9
#135Mistral Large 3 (no thinking) · open weights562 ±3861.8$0.00990.7 min57.062.965.672.758.162.464.153.362.4

Effort Ladder for Mistral Medium 3.5

Only Mistral Medium 3.5 ran at more than one thinking level here. Its page shows how the score, cost and time change from one level to the next.

Mistral Large 4 Against Other Mistral Models

Each comparison lists every thinking level of both models, with Writing Elo, cost and time side by side. The line under each pair gives the Writing Elo gap between the two best configurations. A pair outside the written match-ups opens in the comparison tool.

← All comparisons