Vals AI Multimodal Index: leaderboard
A composite from Vals AI rather than a single task, combining finance document and finance agent benchmarks, coding benchmarks including SWE-bench Verified and Terminal-Bench, and SAGE, which asks models to grade handwritten student mathematics. Each sector is weighted by its share of United States economic output and the weighted total is normalised. Higher is better.
Metric: Accuracy (%). Source: www.vals.ai. Saturation forecast: Around May 2027. 33 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 (Max) | 74.15 | #16 (Claude Fable 5) |
| 2 | Claude Opus 5 | 73.9 | #12 |
| 3 | Kimi K3 | 73.42 | #30 |
| 4 | GPT-5.6 Sol (Max) | 72.64 | #18 (GPT-5.6 Sol) |
| 5 | Claude Opus 4.8 (Max) | 70.89 | #35 (Claude Opus 4.8) |
| 6 | Muse Spark 1.2 | 69.8 | #64 |
| 7 | GPT-5.6 Luna (Max) | 69.06 | #72 (GPT-5.6 Luna) |
| 8 | Claude Sonnet 5 (Max) | 68.83 | #57 (Claude Sonnet 5) |
| 9 | GPT-5.5 (xHigh) | 68.07 | #29 (GPT-5.5) |
| 10 | Claude Opus 4.7 (High) | 67.36 | #50 (Claude Opus 4.7) |
| 11 | Muse Spark 1.1 (xHigh) | 66.74 | #54 (Muse Spark 1.1) |
| 12 | Qwen 3.8 Max | 65.39 | #32 |
| 13 | Gemini 3.6 Flash (High) | 65.08 | #51 (Gemini 3.6 Flash) |
| 14 | GPT-5.6 Terra (xHigh) | 65.07 | #40 (GPT-5.6 Terra) |
| 15 | Grok 4.5 (High) | 63.42 | #49 (Grok 4.5) |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=vals-ai-multimodal-index · How It Works · Data refreshed daily, snapshot 2026-10-23.