Benchmarks / Composite
Vals RSI Index v1.1: Post-training: Finance Agent
Can a model do the research that builds the next model? Autonomous LLM R&D scored against frozen published, human, and frontier-model references. Published Post-training: Finance Agent score in percentage points.
Vals RSI Index v1.1: Post-training: Finance Agent
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Opus 5.5 | max | 35.1% |
| 2 | Opus 5 | max | 31.8% |
| 3 | Fable 5.1 | max | 28.0% |
| 4 | GPT-6 Sol | max | 16.3% |
| 5 | Gemini 4 | max | 15.4% |
| 6 | GPT-6 Astra | max | 7.8% |
| 7 | Kimi K2.5 | max | 0.0% |
| 8 | Muse Spark 1.3 | max | 0.0% |
| 9 | GLM 5.3 | max | 0.0% |
| 10 | GPT-5.2 | xhigh | 0.0% |
| 11 | Qwen 3.8 Max | max | 0.0% |
| 12 | Grok 4.6 | max | 0.0% |
| 13 | GPT-6.1 Sol | max | 0.0% |
| 14 | Gemini 3.7 Flash | max | 0.0% |
| 15 | Opus 4.5 | high | 0.0% |
| 16 | Kimi K3 | max | 0.0% |
| 17 | Grok 4.7 | max | 0.0% |
| 18 | Gemini 3.8 Flash | max | 0.0% |
| 19 | GPT-5.4 | xhigh | 0.0% |
| 20 | GPT-5.6 Sol | max | 0.0% |
| 21 | Opus 4.8 | max | 0.0% |
| 22 | Fable 5 | max | 0.0% |
| 23 | GPT-5.5 | xhigh | 0.0% |
| 24 | Opus 4.7 | max | 0.0% |
| 25 | Wonder Full | max | 0.0% |