Electricity Bench

2026-W38: Vibe coding

Week 38 of 2026 graded 13 coding agents on vibe coding. GPT-6 Astra leads with C+. The biggest move since W37 2026 is GPT-5.6 Luna, F to C-. Grades compare only within the same suite version, and plan usage is each agent's share of its own subscription. Every agent ran the same suite cut headless through its own tools.Measured

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

Published suite snapshot for 13 graded agents, current through Sep 14, 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
GPT-6 Astracodex-cliC+74.2%level 3 · partial level 52m 29s100%0.2%/1wChatGPT Pro 5x
Claude Fable 5.1claude-codeC66.5%level 2 · partial level 53m 47s100%2.2%/5hClaude Max 20x
Gemini 3.8 Flashantigravity-cliC-60.0%level 1 · partial level 57m 35s100%1.0%/1wGoogle AI Pro
GPT-5.6 Lunacodex-cliC-60.0%level 1 · partial level 53m 3s100%0%/1wChatGPT Pro 5x
GPT-5.6 Solcodex-cliC-56.1%level 2 · partial level 54m 33s100%0.4%/1wChatGPT Pro 5x
Claude Opus 5claude-codeD+53.5%level 3 · partial level 55m 39s100%1%/5hClaude Max 20x
Grok 4.6grok-cliD+53.5%level 3 · partial level 511m 13s100%0.4%/1wSuperGrok
Kimi K3 256Kkimi-cliD+52.3%level 1 · partial level 53m 57s100%11.6%/5hKimi Allegretto
GPT-5.6 Terracodex-cliD-41.3%none2m 51s100%0%/1wChatGPT Pro 5x
Claude Sonnet 5claude-codeF32.9%level 1 · partial level 54m 19s100%0.2%/5hClaude Max 20x
Gemini 3.1 Proantigravity-cliF31.6%level 1 · partial level 55m 45s100%1.0%/1wGoogle AI Pro
Composer 2.5cursor-cliF27.1%none1m 56s100%0.1%/1mCursor Pro
Claude Haiku 4.5claude-codeF21.3%none3m 44s100%0.8%/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.