AgentWorldBench (Terminal subset) — Distinct domain subset of AgentWorldBench
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.8 | Anthropic | 59.2 | 1 | 2026-06-24 |
| 2 | Qwen AgentWorld 397B A17B | Alibaba | 57.7 | 1 | 2026-06-24 |
| 3 | Claude Opus 4.6 | Anthropic | 57.5 | 1 | 2026-06-24 |
| 4 | Claude Sonnet 4.6 | Anthropic | 57 | 1 | 2026-06-24 |
| 5 | Qwen3.5 397B A17B | Alibaba | 55.3 | 1 | 2026-06-24 |
| 6 | Qwen AgentWorld 35B A3B | Alibaba | 54 | 1 | 2026-06-24 |
| 7 | GPT-5.4 | OpenAI | 53.7 | 1 | 2026-06-24 |
| 8 | Kimi K2.6 | Moonshot | 52.5 | 1 | 2026-06-24 |
| 9 | Gemini 3.1 Pro | 52.5 | 1 | 2026-06-24 | |
| 10 | DeepSeek-V4-Pro | DeepSeek | 51.3 | 1 | 2026-06-24 |
| 11 | Qwen3.6 Plus | Alibaba | 50.6 | 1 | 2026-06-24 |
| 12 | GLM-5.1 | Z.ai | 47.3 | 1 | 2026-06-24 |
| 13 | Qwen3.5 35B A3B | Alibaba | 46.1 | 1 | 2026-06-24 |
| 14 | MiniMax M2.7 | MiniMax | 41.6 | 1 | 2026-06-24 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.