AgenticRL-Bench Verifiable Head-to-head Elo: leaderboard
AgenticRL-Bench Verifiable Head-to-head Elo — 19verifiable tasks×3attempts, headless one-shot8h, pass@1 solved/57. Three open-ended4×H200 tasks10–16h, means of3counted attempts in their native units. Elo is native BT within-task comparisons, not accuracy. Exact model/harness/effort retained; no best-harness pooling. Metric: Head-to-head Elo. Source: https://agenticrl-bench.env-lab.ai/leaderboard
Metric: Head-to-head Elo. Source: agenticrl-bench.env-lab.ai. Saturation forecast: Rough model projection: around 2026. 7 models tracked. Updated 16 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Claude Opus 5.5 | (Max) (AgenticRL run: harness=Claude Code) | 1132 | #5 |
| 2 | Claude Fable 5.2 | (Max) (AgenticRL run: harness=Claude Code) | 1051 | #1 |
| 3 | Gemini 4 Carbon | (High) (AgenticRL run: harness=Antigravity) | 1033 | #2 |
| 4 | GPT-6 Astra | (Max) (AgenticRL run: harness=Codex) | 983 | #6 |
| 5 | Gemini 4 Argon | (High) (AgenticRL run: harness=Antigravity) | 980 | #12 |
| 6 | Muse Wonder Full 1.0 | (Xhigh) (AgenticRL run: harness=Muse Code) | 925 | |
| 7 | Grok 4.7 | (Xhigh) (AgenticRL run: harness=Grok Build) | 897 | #47 |
Interactive version: theaggregate.ai/benchmark?slug=agenticrl-bench-verifiable-head-to-head-elo · How It Works · Data refreshed daily, snapshot 2026-11-16.