Benchmark
Vals Vibe Code Bench 1-100 v1.1
Coding benchmarks compare models on programming tasks. Scores are percentages.
- Models measured
- 10
- Scores recorded
- 10
- Sources
- 1
- IQ weight
- 2.72
Score distribution
- 33%
- Mean
- 29.8%
- Median
- 11.1%
- Std dev
- 24.5%
- 10th percentile
- 49.7%
- 90th percentile
- 19.8%
- Min
This benchmark has a weight of 2.72 for chat models, measured over 10 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | GPT-6.2 Astra· Vals AI | 55.2% |
| 2 | Gemini 4 Carbon· Vals AI | 49.7% |
| 3 | GPT-6.2 Sol· Vals AI | 40.9% |
| 4 | Claude Opus 5.5· Vals AI | 31% |
| 5 | GLM 5.3· Vals AI | 29.8% |
| 6 | MiMo V2.6 Pro· Vals AI | 26.4% |
| 7 | Grok 4.7· Vals AI | 26.4% |
| 8 | DeepSeek V4.1 Flash 1017· Vals AI | 26.2% |
| 9 | MiMo V2.6 Flash· Vals AI | 24.5% |
| 10 | MiniMax M3.1 Flash· Vals AI | 19.8% |