There is no fixed SKU. Each line below is a production capability of the RSI engine, and customers take the slices they need by subject, level, difficulty and format. Every line has an inspectable sample.
Graduate & Competition Mathematics
Research-grade and elite-competition problems with rigorous, verifiable solutions.
1M+problems→Focus line→ 100MUndergraduate STEM Reasoning
Mathematics, physics, chemistry, biology and computer science at university level and above.
40M+problems→Focus line→ 10MMultimodal Mathematics
University-level problems where the figure is part of the problem: graphs, diagrams, geometry.
1M+image-text problems→AvailableCoding & Office Agent Trajectories
Complete multi-turn agent runs: software-engineering tasks and long-horizon office work.
100K+trajectories per track→AvailableExpert Long-Document Reasoning
Finance, legal and healthcare tasks on real long documents, targeting known model failures.
10K+RL tasks→AvailableMultilingual Speech for ASR & TTS
Precisely aligned speech and text across major and low-resource languages and dialects.
10M+hours→In developmentSynthetic Pre-training Corpus
Textbook-style, long-form reasoning text produced by the RSI engine, for next-generation pre-training.
In development→Compare the lines.
| Line | Scale | Status | Description |
|---|---|---|---|
| Graduate & Competition Mathematics | 1M+ problems | Focus | Research-grade and elite-competition problems with rigorous, verifiable solutions. |
| Undergraduate STEM Reasoning | 40M+ problems → 100M | Focus | Mathematics, physics, chemistry, biology and computer science at university level and above. |
| Multimodal Mathematics | 1M+ image-text problems → 10M | Focus | University-level problems where the figure is part of the problem: graphs, diagrams, geometry. |
| Coding & Office Agent Trajectories | 100K+ trajectories per track | Available | Complete multi-turn agent runs: software-engineering tasks and long-horizon office work. |
| Expert Long-Document Reasoning | 10K+ RL tasks | Available | Finance, legal and healthcare tasks on real long documents, targeting known model failures. |
| Multilingual Speech for ASR & TTS | 10M+ hours | Available | Precisely aligned speech and text across major and low-resource languages and dialects. |
| Synthetic Pre-training Corpus | In development | In development | Textbook-style, long-form reasoning text produced by the RSI engine, for next-generation pre-training. |
Go deeper with us.
Tell us the capability you are training for. We will scope a slice by subject, difficulty and format, and send a pilot sample for your own evaluation.