deepseek-v4-pro vs grok-4.20-0309-reasoning

Capability for the money

grok-4.20-0309-reasoning is 5.5× dearer for 6.4 points on following instructions (IFBench) — worth it only if that is your bottleneck.

1. Following instructions

   extraction, classification, sticking to a format

   IFBench
   deepseek-v4-pro76.5%

   grok-4.20-0309-reasoning82.9%

   grok-4.20-0309-reasoning +6.4

2. Tool use

   multi-step agent work against real tools

   τ²-bench
   deepseek-v4-pro96.2%

   grok-4.20-0309-reasoning96.5%

   grok-4.20-0309-reasoning +0.3

3. Hard reasoning

   graduate-level science and multi-step analysis

   GPQA Diamond
   deepseek-v4-pro88.8%

   grok-4.20-0309-reasoning88.5%

   deepseek-v4-pro +0.3

Averaged over the benchmarks both models sat, deepseek-v4-pro returns 160.3 points of pass rate per dollar of blended price against 29.8 — 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. No published coding score for one of these models, so that lens is omitted. And a benchmark is not your workload — it narrows the shortlist, it does not pick for you.