STT-Arena: leaderboard
Large language models (LLMs) deployed in real-world agentic applications must be capable of replanning and adapting when mid-task disruptions invalidate their prior decisions. Existing dynamic benchmarks primarily measure whether LLMs can detect temporal changes in a timely manner, leaving the complementary challenge of adaptive replanning under spatio-temporal dynamics largely unexplored. We intr
Metric: Overall (self-reported). Source: benchmarklist.com. Saturation forecast: Around July 2027. 23 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 35.39 | #63 |
| 2 | GPT-5.4 | 34.51 | #79 |
| 3 | Gemini 3.1 Pro (Preview) | 33.19 | #58 |
| 4 | Claude Sonnet 4.6 | 32.74 | #90 |
| 5 | DeepSeek V3.2 | 32.16 | #205 |
| 6 | Qwen 3.6 Plus | 31.42 | #120 |
| 7 | GLM-5.1 | 27.31 | #97 |
| 8 | DeepSeek V4 Pro | 26.29 | #102 |
| 9 | GLM-5 | 24.23 | #141 |
| 10 | MiniMax-M2.5 | 21.88 | #303 |
| 11 | MiniMax-M2.7 | 21.29 | #258 |
| 12 | Gemini 2.5 Flash | 20.11 | #245 |
| 13 | Qwen 3.5 9B | 19.53 | #379 |
| 14 | Llama 3.3 70B | 17.18 | #579 |
| 15 | Qwen 3 8B | 14.83 | #674 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=stt-arena · How It Works · Data refreshed daily, snapshot 2026-10-20.