Benchmarks · coding only
The Vibe Coding Index
A full benchmarking deck for AI coding models — quality, price, speed and latency in one place. The Vibe Father runs them all, so we've no reason to pick a favourite.
Models tracked
22
Coding suites
3
in the index
Blended price range
$0.53–$20
per 1M tokens
Data snapshot
Jul 9, 2026
verified + cited
👑
⚡
🪙
⏱
Highest quality
Claude Fable 5
89.9 VCI
Fastest output
Qwen3.7 Max
204 tok/s
Cheapest
MiniMax M3
$0.53 /M blended
Snappiest
DeepSeek V4 Pro
1.87s to first token
✓ Verified data — every figure is cited. See sources.
The composite
Top 10 by Vibe Coding Index
Five coding suites, weighted and normalised to 0–100. Longer bar is better.
Cost × quality
Quality vs price — the efficient frontier
Index against blended token cost (log scale). Up and to the LEFT is better: more quality per dollar.
Speed × quality
Quality vs output speed
Index against generation throughput. Up and to the RIGHT is better: smart and fast.
Blended price — $/M, low → high
all →
GPT-5.5
—
MiniMax M2.1
—
DeepSeek V3.2 (Thinking)
—
GLM 5.2
—
Claude Opus 4.7
—
Kimi K2.6
—
Gemini 3 Pro
—
Grok 4.20
—
Output speed — tok/s, high → low
all →
Qwen3.7 Max
204
Gemini 3.5 Flash
167
Gemini 3.1 Pro
147
Claude Haiku 4.5
96
MiniMax M3
95
Grok 4.5
91
Claude Sonnet 5
89
GPT-5.3 Codex
87
Don't pick the top of the chart. Run the whole chart.
The Vibe Father orchestrates every model here side by side — best model wins each task, on your keys.
Get The Vibe Father