K2 Horizon FTPO Card - IFBench (loose): leaderboard
IFBench's out-of-distribution verifiable instructions, scored with the loose criterion. Scores from IFM's K2 Horizon 3.7B model card at revision 4525090a, whose anti-doom-loop table evaluates the 3.7B and 7B launch checkpoints and the checkpoints IFM promoted on 1 October 2026 after Final Token Preference Optimization (FTPO), with identical settings. Output budget 500k tokens. Metric: accuracy (%).
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around February 2027. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | K2 Horizon 7B | 52 | #714 |
| 2 | K2 Horizon 3.7B | 47.3 | #1075 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=k2-horizon-ftpo-card-ifbench-loose · How It Works · Data refreshed daily, snapshot 2026-11-12.