KAware - External Function: leaderboard
Metric: The knowing-acting score (%, KAS): harmonic mean of knowing accuracy (Jaccard overlap between the tools the agent says the task needs and the gold tool set) and acting accuracy (Jaccard overlap between the tools it actually calls and the gold set), on the external function setting, on the KAware tool-use tasks, where tools are always available but the task may need them (external function, 310 tasks), mix them with parametric knowledge (hybrid composition, 368 tasks) or be solvable from knowledge alone (internal function, 398 tasks); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 235B A22B 2507 (Thinking) | 97.42 |
| 2 | Qwen 3 Max (Preview) | 97.36 |
| 3 | Qwen 3.5 397B A17B | 97.34 |
| 4 | Claude Sonnet 4.5 (Thinking) | 97.3 |
| 5 | Qwen 3.5 397B A17B (Non-reasoning) | 96.62 |
| 6 | DeepSeek V4 Flash (Non-reasoning) | 96 |
| 7 | GPT-5.5 | 95.9 |
| 8 | Claude Sonnet 4.5 | 95.62 |
| 9 | Gemini 3 Pro (Preview) | 91.75 |
| 10 | GPT-4.1 | 91.26 |
| 11 | GPT-4o (2024-11-20) | 90.98 |
| 12 | DeepSeek V3.2 (Non-reasoning) | 90.84 |
| 13 | Gemini 3 Flash (Preview) | 89.61 |
| 14 | O4 Mini (2025-04-16) | 89.22 |
| 15 | GPT-5 | 88.36 |
Interactive version: theaggregate.ai/benchmark?slug=kaware-external-function · How It Works · Data refreshed daily, snapshot 2026-09-29.