← Leaderboard

Mistral · Open weights · tested 17 Nov 2026

Mistral Large 4.0 Max effort

Overall rank#18 of 19
Solves3/1002 first attempt · +1 retry
Rungs proven80/40060 first attempt · +20 retry
Cost to run$104$34.60 per solve
Tokens used743.8M92.1% from cache · 10.6M output
Time per lab25.6 minaverage of 195 attempts

Where it places

Each line is one metric across all 19 ranked configurations. Right is always better; this model is the large dot.

Solveslabs solved, of 100
worsebetter
3#18 of 19
Rungs provenrungs proven, of 400
worsebetter
80#18 of 19
First-attempt solvessolved on the first attempt
worsebetter
2#17 of 19
Cost per solveUSD per solved lab
worsebetter
$34.60#14 of 18

By vulnerability class

How Mistral Large 4.0 (Max) did on each type of vulnerability. Shown as counts: each class has 10 labs, so one lab is 10% of a class.

  • Access control1 of 10 solved17 of 40 rungs
  • Authentication & sessions1 of 10 solved12 of 40 rungs
  • Injection: queries & data0 of 10 solved4 of 40 rungs
  • Code execution0 of 10 solved6 of 40 rungs
  • Server-side access0 of 10 solved2 of 40 rungs
  • Browser-side0 of 10 solved4 of 40 rungs
  • Infrastructure & edge0 of 10 solved8 of 40 rungs
  • Business logic & races1 of 10 solved10 of 40 rungs
  • Secrets, crypto & supply chain0 of 10 solved11 of 40 rungs
  • AI agents0 of 10 solved6 of 40 rungs

Provider performance

Measured per single agent against the provider endpoint used for the run.

Response time
9.34 s median · 32.92 s p95
Output speed
49.2 tok/s
API success rate
98.98%
Input context
39.5K median · 99.1K p95

About this result

Every model runs in the same neutral harness, so results compare the models themselves.

Harness
BB Arena neutral harness
Safety refusals
None

See Mistral Large 4.0 (Max) highlighted on every chart →