Scaling RSI
RL training environments for long-horizon SWE and ML.
Work with us300+ environments ready for training
We’re building training environments for recursive self-improvement: an agent’s ability to help build the next version of itself.
We believe high-quality RSI data is the bottleneck. Our environments cover improving sample efficiency and model architectures, debugging training runs, and making open-ended research decisions.
We have partnered with several frontier labs and are committed to increasing sovereign intelligence worldwide.
Results
Models 4 of 20
Show models
| Model cohort | Mean reward | Mean cost per task | Avg. output tokens (M) |
|---|---|---|---|
| 0.2489 | $129.28 | 0.602 | |
| 0.2292 | $31.58 | 0.720 | |
| 0.2279 | $22.72 | 0.834 | |
| 0.1906 | $11.95 | 0.439 | |
| 0.1755 | $29.59 | 0.368 | |
| 0.1740 | $35.20 | 1.132 | |
| 0.1653 | $49.17 | 0.189 | |
| 0.1580 | — | 0.159 | |
| 0.1409 | $21.94 | 0.435 | |
| 0.1130 | $2.03 | 0.433 | |
| 0.0903 | $0.92 | 0.237 | |
| 0.0898 | $10.06 | 0.265 | |
| 0.0778 | $0.95 | 0.327 | |
| 0.0746 | $17.38 | 0.172 | |
| 0.0727 | $7.15 | 0.121 | |
| 0.0712 | $5.27 | 0.116 | |
| 0.0661 | $1.55 | 0.255 | |
| 0.0516 | $10.64 | 0.126 | |
| 0.0235 | $3.91 | 0.047 | |
| 0.0179 | $4.93 | 0.210 |
Choose a model to see its results.
4 models shown.