TKFQA - Reasoning-Chain Accuracy (Table-Text-KG Order): leaderboard
Metric: Reasoning-chain accuracy (%; entity-level agreement of the generated reasoning chain with the annotated counterfactual chain, for counterfactual multi-hop questions grounded in a table, text passages and a knowledge graph, 1,012 test items; contexts given in table, text, knowledge-graph order; mean of three runs). Source: arxiv.org. Saturation forecast: Around December 2026. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O3 | 79.05 |
| 2 | Gemini 2.5 Flash | 77.93 |
| 3 | DeepSeek V3.2 | 69.2 |
| 4 | O4 Mini | 68.18 |
| 5 | DeepSeek V4 Flash | 65.26 |
| 6 | Kimi K2.5 | 64.04 |
| 7 | MiniMax-M2.7 | 55.34 |
| 8 | GPT-5 | 51.52 |
| 9 | Qwen 3 30B A3B | 43.31 |
| 10 | Gemini 2.5 Flash Lite | 37.25 |
| 11 | Grok 4.3 | 32.61 |
| 12 | GPT-4.1 Mini | 29.59 |
| 13 | Qwen 3 8B | 20.14 |
| 14 | Gemma 3 12B | 19.95 |
Interactive version: theaggregate.ai/benchmark?slug=tkfqa-reasoning-chain-accuracy-table-text-kg-order · How It Works · Data refreshed daily, snapshot 2026-09-26.