Electricity Bench

claude-code + claude-sonnet-5 vs codex-cli + gpt-5.6-sol

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

claude-cliclaude-code + claude-sonnet-5Solves up to tier 5overall pendingcodex-clicodex-cli + gpt-5.6-solSolves up to tier 5 at 3.6% of ChatGPT Plus per run · 27.8 runs/weekoverall pending
real-world@v3FvsD
Dimensionclaude-code + claude-sonnet-5codex-cli + gpt-5.6-solDelta
Capability33%47%13% · codex-cli + gpt-5.6-sol
Median latency228312 ms211436 ms16876 ms · codex-cli + gpt-5.6-sol
Reliability100%100%even
Error rate0%0%even

claude-code + claude-sonnet-5: plan cost not measured. codex-cli + gpt-5.6-sol: 3.6% of ChatGPT Plus per run, about 27.8 runs per week. 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.