GTA-Workflow: leaderboard
Root success on GTA-Workflow, the open-ended tier of GTA-2 (2026): 132 long-horizon productivity workflows from real user queries with deployed tools and multimodal inputs, run end to end in a ReAct agent framework and judged by GPT-5.2 against recursive checkpoint rubrics: the share of tasks whose final deliverable scores above 7 of 10 at the root checkpoint. Percent; higher is better.
Metric: Root success rate (%): share of tasks whose root checkpoint score exceeds 7 of 10 on GTA-Workflow, the open-ended long-horizon workflow tier of GTA-2 (132 real-user tasks with deployed tools and multimodal inputs), run end to end in the default Lagent ReAct framework; deliverables judged by GPT-5.2 against recursive checkpoint rubrics; higher is better. Source: arxiv.org. Saturation forecast: Around November 2027. 13 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 14.39 | #276 |
| 2 | GPT-5 | 11.36 | #178 |
| 3 | Qwen 3 235B A22B | 10.61 | #628 |
| 4 | Llama 4 Scout | 10.61 | #1818 |
| 5 | Grok 4 | 9.85 | #319 |
| 6 | Claude Sonnet 4.5 | 9.09 | #261 |
| 7 | DeepSeek V3.2 | 9.09 | #419 |
| 8 | Kimi K2 | 8.33 | #443 |
| 9 | Llama 3.1 70B Instruct | 0.76 | #1571 |
| 10 | Llama 3.2 3B Instruct | 0.76 | #7241 |
| 11 | Qwen 3 30B A3B | 0.76 | #1156 |
| 12 | Llama 3.1 8B Instruct | 0 | #4600 |
| 13 | Qwen 3 8B | 0 | #1873 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=gta-workflow · How It Works · Data refreshed daily, snapshot 2026-11-21.