Benchmark
Vals Finance Agent (v2)
Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.
- Models measured
- 78
- Scores recorded
- 78
- Sources
- 2
- IQ weight
- 2.83
Score distribution
- 48.4%
- Mean
- 51.9%
- Median
- 11.6%
- Std dev
- 31%
- 10th percentile
- 59%
- 90th percentile
- 9%
- Min
This benchmark has a weight of 2.83 for chat models, measured over 78 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 | Qwen3.7 Max· Vals AI | 47.8% |
| 52 | Gemini 3.5 Flash Lite· Vals AI | 47.4% |
| 53 | Inkling FP4· Vals AI | 46.6% |
| 54 | GPT-5.4 mini· Vals AI | 45.4% |
| 55 | Kimi K2.6· Vals AI | 44.9% |
| 56 | GLM 5.1· Vals AI | 44.8% |
| 57 | DeepSeek V4 Pro· Vals AI | 44.1% |
| 58 | Gemini 3.1 Pro Preview· Vals AI | 43% |
| 59 | Gemini 3 Flash· Vals AI | 42.6% |
| 60 | MiMo V2.5 Pro· Vals AI | 41.5% |
| 61 | Inkling Small· Vals AI | 41.3% |
| 62 | Qwen3.6 Plus· Vals AI | 40.8% |
| 63 | Qwen3.7 Plus· Vals AI | 38.2% |
| 64 | GPT 5.4 Nano· Vals AI | 38.2% |
| 65 | Grok 4.3· Vals AI | 37.7% |
| 66 | Nemotron 3 Ultra· Vals AI | 37.7% |
| 67 | MiMo V2.5· Vals AI | 36.7% |
| 68 | Kimi K2.5thinking· Vals AI | 35.8% |
| 69 | Mistral Medium 3.5· Vals AI | 32.1% |
| 70 | Claude Haiku 4.5thinking· Vals AI | 31% |
| 71 | Gemini 3.1 Flash Lite· Vals AI | 30% |
| 72 | Grok 4.20· Vals AI | 28.5% |
| 73 | MiniMax M2.7· Vals AI | 27.9% |
| 74 | Laguna M.1· Vals AI | 25% |
| 75 | Mercury 2.5· Vals AI | 18.9% |
| 76 | Nemotron 3.5 Lightning 30B A3B· Vals AI | 18.5% |
| 77 | Laguna XS.2· Vals AI | 15.6% |
| 78 | Command A Plus 05 2026· Vals AI | 9% |
Showing 51–78 of 78 results.
Data credits: Data by LLM Stats