Thinking Levels · Mistral
Last updated , from board data generated at 17:52 UTC.
This page compares Mistral Medium 3.5 at 2 reasoning-effort settings, each tested on the same 10 scripts: the Writing Score (the 0 to 100 consensus of the judges and code), the Writing Elo (a rating from pairwise comparisons of those scores), and the cost and time of each setting. As a model, Mistral Medium 3.5 ranks #73 of 87 models through its best setting. The other ranks on this page are configuration ranks: one setting's place among all 151 configurations.
Reasoning gives Mistral Medium 3.5 a small gain that isn't worth its price. With reasoning on, it gains 24 Elo and moves from #136 to #134, with overlapping ranges; that alone doesn't settle their order. Meanwhile each script costs 6.0 times as much, going by estimated prices for both settings, and takes 2.1 minutes instead of 0.2.
I'd keep reasoning off. It trades one weakness for others: scripts land much closer to the target length, but the craft, slop and visual cue scores all get worse, and the overall score ends up about the same. Paying that much more for a reshuffle makes no sense to me.
Neither setting is one I'd use for scripts. If you want to stay with Mistral, Mistral Large 4 (high) ranks #77 for $0.089 a script, still below the estimated price of reasoning mode here. It's slow, though, at 8.7 minutes a script.
Each point is one effort setting of Mistral Medium 3.5, joined from the lowest setting to the highest. Higher and further left is better: a higher score for less money or less time.
On a small screen, Table lists the same numbers as rows.
| Setting | Writing Elo | Step | Writing Score | Cost / Script | Time / Script | Reasoning Tokens | Configuration Rank |
|---|---|---|---|---|---|---|---|
| no thinking (provider default) | 553 ±41 | - | 61.3 | $0.033* | 0.2 min | - | #136 |
| high | 577 ±60 | +24 | 61.1 | $0.201* | 2.1 min | 19,111 | #134 |
The 95% Writing Elo ranges overlap: Mistral Medium 3.5 (no thinking) runs from 510 to 593, and Mistral Medium 3.5 (high) from 520 to 639.
Writing Elo comes from comparing every pair of models on every script, which puts these settings on the same scale as the leaderboard. Step is the Writing Elo change from the provider default. Cost is one full script at API list prices, and time is the average for one accepted script, retries included. * marks an estimated price. The thinking-levels page covers the other models, and the Mistral family page lists every Mistral model side by side.