IFM K2-Horizon-375B Card - APEX-Agents (Text-Only Subset): leaderboard
APEX-Agents evaluates agents on long-horizon professional workflows. IFM ran this series for its K2-Horizon-375B-A23B model card on a subset of text-only tasks, so values are not comparable with full-benchmark results. Pass@1 in percent, higher is better; K2-Horizon runs at high reasoning effort and the other columns carry their own effort labels.
Metric: Pass@1 (%). Source: huggingface.co. Saturation forecast: Around June 2028. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 5 (Max) | 31.7 | #65 (Claude Sonnet 5) |
| 2 | GPT-5.6 Luna (Max) | 28.6 | #79 (GPT-5.6 Luna) |
| 3 | GLM-5.2 (Max) | 26.9 | #80 (GLM-5.2) |
| 4 | GPT-5.6 Terra (High) | 25.4 | #40 (GPT-5.6 Terra) |
| 5 | MiniMax-M3 | 23.8 | #151 |
| 6 | Inkling (xHigh) | 19 | #144 (Inkling) |
| 7 | Nemotron 3 Ultra | 9 | #245 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ifm-k2-horizon-375b-card-apex-agents-text-only-subset · How It Works · Data refreshed daily, snapshot 2026-11-05.