Benchmark
Vals Terminal-Bench 4.0
Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.
- Models measured
- 65
- Scores recorded
- 65
- Sources
- 1
- IQ weight
- 2.05
Score distribution
- 28.6%
- Mean
- 24.2%
- Median
- 21.2%
- Std dev
- 2.5%
- 10th percentile
- 60.6%
- 90th percentile
- 0%
- Min
This benchmark has a weight of 2.05 for chat models, measured over 65 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 |
|---|---|---|
| 51 | Grok 4.5· Vals AI | 8.6% |
| 52 | Hy4 preview· Vals AI | 8.1% |
| 53 | Muse Spark Open· Vals AI | 6.6% |
| 54 | Gemini 3.5 Flash· Vals AI | 6.1% |
| 55 | Muse Spark 1.2· Vals AI | 6.1% |
| 56 | Beam 1.0 501B A23B· Vals AI | 4% |
| 57 | Mercury 3· Vals AI | 3% |
| 58 | Gemini 3.1 Pro Preview· Vals AI | 2.5% |
| 59 | GPT-5.4 mini· Vals AI | 2.5% |
| 60 | MiniMax M3· Vals AI | 1% |
| 61 | Inkling FP4· Vals AI | 0.505% |
| 62 | MiMo V2.5 Pro· Vals AI | 0.505% |
| 63 | Inkling Small· Vals AI | 0% |
| 64 | Mercury 2.5· Vals AI | 0% |
| 65 | NVIDIA Nemotron 3.5 Super VL 120B· Vals AI | 0% |
Showing 51–65 of 65 results.