Taste-Bench - Parallel Research: leaderboard
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around December 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 | 64.6 | #11 |
| 2 | GLM-5.2 (Max) | 60.4 | #70 (GLM-5.2) |
| 3 | GPT-5.6 Terra | 58.3 | #40 |
| 4 | MiniMax-M3 | 56.2 | #130 |
| 5 | GPT-5.5 (xHigh) | 56.2 | #29 (GPT-5.5) |
| 6 | GPT-5.6 Sol (xHigh) | 56.2 | #18 (GPT-5.6 Sol) |
| 7 | Grok 4.5 (High) | 56.2 | #46 (Grok 4.5) |
| 8 | Claude Sonnet 5 (Thinking) | 56.2 | #56 (Claude Sonnet 5) |
| 9 | GPT-5.4 Mini (High) | 54.2 | #206 (GPT-5.4 Mini) |
| 10 | GPT-5.6 Luna | 50 | #67 |
| 11 | DeepSeek V4 Flash (Max) | 50 | #134 (DeepSeek V4 Flash) |
| 12 | GPT-5.4 Nano (High) | 43.8 | #320 (GPT-5.4 Nano) |
| 13 | Mistral Medium 3.5 (High) | 41.7 | #338 (Mistral Medium 3.5) |
| 14 | Grok 4.20 (Reasoning) | 8.3 | #169 (Grok 4.20) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=taste-bench-parallel-research · How It Works · Data refreshed daily, snapshot 2026-10-13.