Electricity Bench

Grok 4.6 vs Grok 4.7

Grok 4.6 was superseded by Grok 4.7. The grades below are the last ones measured, so read them as history rather than as this week's board. See the current roster.

Both grade C- overall. The two cost about the same per task. By suite, real-world issues is even at D and spec planning is even at C. Both agents run the same private suite cut headless through their own tools, each on its own subscription.Measured

Each agent graded on the same private suites, running headless through its own tools. Advantages are oriented so a win is a win, whichever way the metric runs.

Verdict: Grok 4.7 vs Grok 4.6

Both grade C- overall. By suite: real-world issues is even at D (47% vs 47% capability); spec planning is even at C (66% vs 68% capability); Grok 4.7 takes vibe coding, C to C- (64% vs 61% capability). Grok 4.6 used 13% of SuperGrok Plus's weekly limit per pass, $0.157 per task at the listed price. Grok 4.7 used 12% of SuperGrok Plus's weekly limit per pass, $0.145 per task at the listed price. Per task the two cost about the same ($0.157 vs $0.145). Grok 4.7 is faster on real-world issues: a median task takes 6m 57s against 7m 10s. Both finished every task on it.

Changed since W38: Grok 4.6: real-world issues C- to D, real-world issues usage 4% to 3% of the weekly limit, vibe coding C+ to C-; Grok 4.7 has one graded week under this version (W39).

Grok 4.6Clears tier 4 (partial tier 5) at 3.0% of SuperGrok Plus per runC-OVERALLGrok 4.7Clears tier 4 (partial tier 5) at 5.0% of SuperGrok Plus per runC-OVERALL
real-world@v3DvsD
DimensionGrok 4.6Grok 4.7Advantage
Capability47%47%Even
Median task time7m 10s6m 57s13s faster (Grok 4.7 leads)
Approx. task cost$0.046 / task$0.077 / task$0.031 / task lower (Grok 4.6 leads)

Grok 4.6: 3.0% of SuperGrok Plus (3.0% / 1w) per run. Grok 4.7: 5.0% of SuperGrok Plus (5.0% / 1w) per run. Each figure is a share of that agent's own plan, so there is no delta to draw.

spec-planning@v1CvsC
DimensionGrok 4.6Grok 4.7Advantage
Capability66%68%+3% (Grok 4.7 leads)
Median task time45m 26s37m 24s8m 1s faster (Grok 4.7 leads)
Approx. task cost$0.575 / task$0.402 / task$0.172 / task lower (Grok 4.7 leads)

Grok 4.6: 10.0% of SuperGrok Plus (10.0% / 1w) per run. Grok 4.7: 7.0% of SuperGrok Plus (7.0% / 1w) per run. Each figure is a share of that agent's own plan, so there is no delta to draw.

vibe-coding@v1C-vsC
DimensionGrok 4.6Grok 4.7Advantage
Capability61%64%+3% (Grok 4.7 leads)
Median task time11m 14s11m 10s3s faster (Grok 4.7 leads)
Approx. task cost<$0.230 / task$0.092 / task$0.138 / task lower (Grok 4.7 leads)

Grok 4.6: under 1% of SuperGrok Plus per good solve. Grok 4.7: 0.4% of SuperGrok Plus (0.4% / 1w) per good solve. Each figure is a share of that agent's own plan, so there is no delta to draw.

Grades and deltas are comparable only within a suite version. See methodology.

Frequently asked questions

Which is better for coding, Grok 4.6 or Grok 4.7?

They are even on our ladder: both clear tier 4 of 5 at 47% capability on real-world@v3, our private suite of real-world coding problems. Latest measurement 2026-09-20.

Which is cheaper per task, Grok 4.6 or Grok 4.7?

Grok 4.6 used 13% of SuperGrok Plus's weekly limit per pass, $0.157 per task at the listed price. Grok 4.7 used 12% of SuperGrok Plus's weekly limit per pass, $0.145 per task at the listed price. Per task the two cost about the same ($0.157 vs $0.145).

What do Grok 4.6 and Grok 4.7 cost a month?

Grok 4.6 runs on SuperGrok Plus at $100/month (public price as of 2026-09-14). Grok 4.7 runs on SuperGrok Plus at $100/month (public price as of 2026-09-14).