HEAD TO HEAD
SAGEA: mira and Rizzo AI Academy: rizzo-flow, compared on what they decide, where they run, what they cost and what their publishers report.
| Property | sagea/mira | rizzo-ai-academy/rizzo-flow |
|---|---|---|
| Summary | Small open typed-decision model with calibrated probabilities, first-class Nepali coverage. | A local, Jev-compatible decision model from Rizzo AI Academy. XHToken's Spark-X2.5-4B with a merged typed-decisions LoRA, run on llama.cpp; yes/no, choice and score questions share one prefill of the state and are read from the answer-letter logits. No text is generated. |
| Decides | choice, score, noul, classify, route | choice, score, noul, classify, route |
| Architecture | mira | rizzo-flow |
| Fine-tuned from | jhu-clsp/mmbert-small | xhtoken/spark-x2.5-4b |
| License | apache-2.0 | apache-2.0 |
| Availability | Open weights | Open weights |
| Hosted by | — | — |
| Input price | — | — |
| Decision accuracy | 61.3% | 64.8% |
| Calibration error | 0.083 | 0.112 |
| Valid action rate | 100.0% | — |
| Median latency | 30 ms | 195 ms |
| p95 latency | 43 ms | — |
| Evaluation suite | s1-decision-bench | LocalLLaMA/typed-decisions test split (400 cases, 2,000 decisions), Q8_0 |
| Latest version | 0.1.5 | 2026.09 |
| Variants | — | lora, LICENSE |
| Size of latest version | 568.8 MB | 22.0 GB |
| Files | 12 | 17 |
| Downloads | 209 | 2 |
| Stars | 1 | 0 |
| Tags | routing, nepali, nepal, calibration, cpu | system-one, spark, lora, gguf, llama-cpp, 4b |
| Updated | Nov 18, 2026 | Oct 7, 2026 |
Figures are from each model’s manifest; accuracy and latency are what the publishers report, on their own suites and hardware. Add a third model.
mira is from SAGEA and rizzo-flow from Rizzo AI Academy. Both have open weights you can download and run. Both answer choice, score, noul, classify and route questions.
They report on different suites — mira 61.3% on s1-decision-bench, rizzo-flow 64.8% on LocalLLaMA/typed-decisions test split (400 cases, 2,000 decisions), Q8_0 — so the numbers do not rank them. Test both on your own labelled examples.
mira: Free (open weights). rizzo-flow: Free (open weights). Open weights cost nothing per call beyond your own hardware.
By their publishers’ figures, mira answers in about 30 ms at the median and rizzo-flow in about 195 ms — measured on different hardware, so treat it as a rough guide.
Yes, both: systemone pull sagea/mira and systemone pull rizzo-ai-academy/rizzo-flow download the weights.