grok-4-1-fast vs mistral-large-3

Capability for the money

mistral-large-3 is 2.7× dearer for 6.6 points on coding (LiveCodeBench) — worth it only if that is your bottleneck.

1. Following instructions

   extraction, classification, sticking to a format

   IFBench
   grok-4-1-fast36.5%

   mistral-large-336.2%

   grok-4-1-fast +0.3

2. Coding

   writing and fixing code

   LiveCodeBench
   grok-4-1-fast39.9%

   mistral-large-346.5%

   mistral-large-3 +6.6

3. Tool use

   multi-step agent work against real tools

   τ²-bench
   grok-4-1-fast63.7%

   mistral-large-324.6%

   grok-4-1-fast +39.1

4. Hard reasoning

   graduate-level science and multi-step analysis

   GPQA Diamond
   grok-4-1-fast63.7%

   mistral-large-368.0%

   mistral-large-3 +4.3

Averaged over the benchmarks both models sat, grok-4-1-fast returns 185.3 points of pass rate per dollar of blended price against 58.4 — the ratio that decides whether the dearer model earns its rate.

Each figure is a pass rate: the share of that test set the model answered correctly, measured by a third party rather than self-reported by the lab. Rows are different tests of different difficulty, so compare a model against the other model, not one row against the next. And a benchmark is not your workload — it narrows the shortlist, it does not pick for you.