Vals Multimodal Index Terminal-Bench 2.1
Indexes · 2026-06-10
This row reports the component score for Terminal-Bench 2.1 within the Vals Multimodal Index composite.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5.6 Sol | 85.8 |
| Claude Opus 5 | 84.6 |
| Kimi K3 | 80.9 |
| Claude Fable 5 | 80.5 |
| GPT-5.6 Luna | 79.0 |
| GPT-5.5 | 76.4 |
| Sonnet 5 | 74.5 |
| Gemini 3.5 Flash | 74.2 |
| Gemini 3.6 Flash | 73.8 |
| GPT-5.6 Terra | 73.4 |
| Opus 4.8 | 71.9 |
| Gemini 3.1 Pro Preview | 70.8 |
| Muse Spark 1.1 | 69.3 |
| Opus 4.7 | 68.5 |
| Grok 4.5 | 67.8 |
| MiMo-V2.5 | 60.7 |
| Sonnet 4.6 | 57.3 |
| GPT-5.4 Mini | 54.7 |
| Gemini 3 Flash Preview | 53.9 |
| Kimi K2.6 | 53.6 |
| MiniMax M3 | 53.6 |
| Qwen3.6 Plus (2026-04-02) | 53.2 |
| Qwen3.7-Plus | 52.8 |
| Grok 4.20 | 44.2 |
| Haiku 4.5 | 43.8 |
| Grok 4.3 | 41.9 |
| GPT-5.4 Nano | 41.6 |
| Mistral Medium 3.5 | 39.0 |
| Gemini 3.1 Flash-Lite Preview | 34.1 |