Model comparisonSo sánh model模型对比
hal-x/thx-01 Details | subquadratic/subq-snap-preview Details | microsoft/microsoft-decision-1 Details | |
|---|---|---|---|
| Overview | |||
| Description | THX-01 is a 322M multilingual decision model from HAL-X AI, trained with reinforcement learning on a strictly proper scoring rule so its probabilities are calibrated. Send a state and typed questions (noul / choice / score) and get an answer with probabilities for each question in one forward pass of about 10 ms — not generated chat text. Post-trained in 18 languages, 100+ supported. Apache 2.0. Call POST /api/v1/decisions (POST /api/v1/systemone is the same handler). | SubQ Snap Preview is Subquadratic's System One decision model for routing, guardrails, triage and other fast decisions. It is in public preview. Send application state and typed questions (noul / choice / score) and get back typed answers with probabilities and confidence — not generated chat text. SubQ documents text input and output plus image input (PNG, JPEG or WebP, up to 4 images). Call POST /api/v1/decisions (POST /api/v1/systemone is the same handler). | Microsoft-Decision-1 is a small Microsoft model built for fast decisions: yes/no, multiple-choice and rating questions, plus rubric-based grading of AI responses and agent actions. Send application state and typed questions (noul / choice / score) and get back a calibrated probability for each fixed answer option, not generated chat text. It has a 32,768-token context window. Call POST /api/v1/decisions (POST /api/v1/systemone is the same handler). |
| Category | Text Generation | Text Generation | Text Generation |
| Context length | 1K | 0 | 33K |
| Providers | 1 | 1 | 2 |
| Pricing (per 1M tokens) | |||
| Input | $0 | $0 | $0.04 |
| Output | $0 | $0 | $0 |
| Capabilities | |||
| Reasoning | – | – | – |
| Tool calling | – | – | – |
| Vision | – | – | |
| Streaming | – | – | – |
| Token activity (30 days) | |||
| Usage |