Electricity Bench

claude-code + claude-opus-5

claude-code harness · Claude Max 5x

plan
Claude Max 5x
price
$100/mo (2026-07-27)
allowance
100 provider-percent per 5h
isolation
quiet-window
usage method
reported-delta

operator Claude Max 5x account on eb-runner, shared with the operator IDE panes; quiet-window until a dedicated benchmark login exists

Solves up to tier 5 on real-world@v3 - the highest difficulty rung with a solved task
12345

Tier 1: localized single-file bug · tier 3: cause in a different module than the symptom · tier 5: problems that took a human hours. A cleared rung solved every task on it; a partial rung solved some.

overall pending Two distinct suite families are required for an overall grade.

Per-suite results

C
real-world@v3
60%
plan cost per run
64.0% of Claude Max 5x
latency med / p95
222605 ms / 795758 ms
reliability
100%
error rate
0%

Run on claude-code 2.1.221 (Claude Code), model claude-opus-5 (--effort medium), Aug 5, 2026.

One run of this suite consumes 64.0% of Claude Max 5x - about 1.6 runs per 5h. Measured per suite run (reported-delta); an absolute cost on this plan, not comparable with another agent's plan. Measured in a declared quiet window on a shared subscription: operator Claude Max 5x account on eb-runner, shared with the operator IDE panes; quiet-window until a dedicated benchmark login exists.

As work per window: about 228 suite runs per month on this plan - 2.28 runs per plan-dollar (at $100/mo, priced 2026-07-27). Work figures compare outputs and are comparable across agents; the raw quota share above is not.

Trajectory

Every scored run of this agent, oldest first. Grades compare only within a suite version, so each version gets its own line; while a suite is being retired both versions run, and the superseded one is dashed. The filled point is the run the scorecard shows today. All agents over time →

real-world · 1 run · latest Aug 5, 2026
Capability
0%50%100%Aug 5claude-code + claude-opus-5 · v3 · Aug 5, 2026 · 60% · grade C0%50%100%Aug 5claude-code + claude-opus-5 · v3 · Aug 5, 2026 · 60% · grade C
Median latency
178084 ms222605 ms267126 msAug 5claude-code + claude-opus-5 · v3 · Aug 5, 2026 · 222605 ms · grade C178084 ms222605 ms267126 msAug 5claude-code + claude-opus-5 · v3 · Aug 5, 2026 · 222605 ms · grade C
Plan cost per run - a share of this agent's own plan, not a cross-agent number
51.2%64.0%76.8%Aug 5claude-code + claude-opus-5 · v3 · Aug 5, 2026 · 64.0% · grade C51.2%64.0%76.8%Aug 5claude-code + claude-opus-5 · v3 · Aug 5, 2026 · 64.0% · grade C

Aug 5, 2026: C on v3

Compare

Transcript excerpts

Outputs are shown only when they can be separated safely from the private suite. Scores and run metadata remain visible when an output is withheld.

real-world@v3
Task#TurnsLatencyOutput
b2r5h8127197950 mswithheldPrivate-suite content
c8g2m4120110687 mswithheldPrivate-suite content
d9b4x6134343053 mswithheldPrivate-suite content
h5n2v7178721258 mswithheldPrivate-suite content
h7w3j5120115907 mswithheldPrivate-suite content
k4f9t7138355597 mswithheldPrivate-suite content
m6t3q8131222605 mswithheldPrivate-suite content
p1x9k4185744461 mswithheldPrivate-suite content
q8k3j6172817123 mswithheldPrivate-suite content
r3v8m5153566157 mswithheldPrivate-suite content
s2j7f4119128876 mswithheldPrivate-suite content
t5s8n2114105745 mswithheldPrivate-suite content
w4j7q2171786603 mswithheldPrivate-suite content
y3p7k1122222454 mswithheldPrivate-suite content
z6q1v9121181069 mswithheldPrivate-suite content