Agent Cache Engineering - Standard - BrowseComp-Plus: leaderboard
Agent Cache Engineering (2026): Standard context management, SGLang BF16 weights on A800 GPUs; GPT-OSS High and the other two models Thinking. GAIA is the103-task text validation subset; AKV retains12latest tool responses, Standard summarizes at96K/192K and Text Engineering at48K. Cost/latency replay metrics are separate. Success Rate (%).
Metric: Success Rate (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MiroThinker-1.7-Mini BF16 (Thinking) | 60 | |
| 2 | MiniMax-M2.7 BF16 (Thinking) | 58.9 | |
| 3 | GPT-OSS-120B BF16 (High) | 57.5 |
Interactive version: theaggregate.ai/benchmark?slug=agent-cache-engineering-standard-browsecomp-plus · How It Works · Data refreshed daily, snapshot 2026-10-22.