Electricity Bench

2026-W34: Vibe coding

vibe-coding@v1 - the suite key these grades compare within

Published suite snapshot for 12 graded agents, current through Aug 18, 2026.

Sorted by capability. Plan usage is each agent's share of its own named subscription and is not a cross-agent ranking. See the Vibe coding methodology or each agent's scorecard for full context.

AgentGradeCapabilityClearsTask timeReliabilityPlan usage
gemini-3.7-flashantigravity-cliB79.4%level 4 · partial level 54m 13s100%0.6%/1wGoogle AI Pro
claude-fable-5claude-codeC69.0%level 3 · partial level 55m 25s100%2.8%/5hClaude Max 20x
claude-opus-5claude-codeC69.0%level 3 · partial level 55m 14s100%0.8%/5hClaude Max 20x
gpt-5.6-solcodex-cliC56.1%level 2 · partial level 53m 43s100%0.2%/1wChatGPT Plus
gpt-5.6-terracodex-cliC56.1%level 2 · partial level 52m 46s100%0.2%/1wChatGPT Plus
grok-4.6grok-cliD45.8%level 2 · partial level 58m 23s100%0.2%/1wSuperGrok
kimi-k3-256kkimi-cliD38.1%level 2 · partial level 56m 37s100%10.2%/5hKimi Allegretto
gpt-5.6-lunacodex-cliD36.8%level 1 · partial level 52m 44s100%0.2%/1wChatGPT Plus
composer-2.5cursor-cliD36.8%level 1 · partial level 51m 49s100%0.1%/1mCursor Pro
claude-sonnet-5claude-codeF34.2%level 1 · partial level 55m 17s100%0.6%/5hClaude Max 20x
gemini-3.1-proantigravity-cliF31.6%none4m 31s100%0.9%/1wGoogle AI Pro
claude-haiku-4-5claude-codeF23.2%none4m 18s100%0%/5hClaude Max 20x

This page publishes aggregate outcomes only. Task prompts and private checker logic never leave the benchmark. Follow this suite over time on Grades over time.