PokeAgent Challenge: leaderboard
Pokemon battling benchmark evaluating AI strategic reasoning, opponent modeling, and decision-making under partial observability in competitive Gen 9 OU battles with extended timers for LLM agents.
Metric: Elo Rating. Source: pokeagentchallenge.com. Saturation forecast: Around May 2027. 33 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 1924 | #58 |
| 2 | GPT-5 | 1906 | #98 |
| 3 | GLM-5 | 1773.4 | #141 |
| 4 | Claude Opus 4.6 | 1670.2 | #63 |
| 5 | Grok 3 Mini | 1598.5 | #281 |
| 6 | Claude Sonnet 4.6 | 1561.8 | #90 |
| 7 | Grok 3 | 1553.4 | #304 |
| 8 | MiniMax-M2.5 | 1543 | #303 |
| 9 | Hermes 4 405B | 1499.6 | #334 |
| 10 | DeepSeek V3 | 1476.8 | #321 |
| 11 | Kimi K2.5 | 1472.5 | #144 |
| 12 | Qwen 3.5 Plus | 1458.3 | #125 |
| 13 | Gemini 3 Flash | 1390.6 | #99 |
| 14 | Gemini 3 Pro | 1359.4 | #80 |
| 15 | Gemini 2.5 Flash | 1232.3 | #245 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=pokeagent-challenge · How It Works · Data refreshed daily, snapshot 2026-10-20.