MOTHRAG claims strong multi-step RAG without GPUs or fine-tuning
MOTHRAG is a system for answering questions that require several evidence steps. Its reader model, embedder, and retrieval judges all run through ordinary pay-per-call APIs. That means it does not require , a local GPU, or proprietary licenses.
Using Llama-3.3-70B as the reader with one shared setup, it scored 78.1 F1 on HotpotQA, 76.3 F1 on 2WikiMultiHopQA, and 50.5 F1 on MuSiQue, for a 68.3 average. The claimed result is within 0.7 points of GPU-heavy systems. The is $0.032 per query, with an economy tier at $0.018 per query that stayed statistically similar on HotpotQA and 2Wiki.
The uses swappable judges to check relevance and sufficiency. Answers are built as proof trees, so each reasoning step can be audited, and the reader, embedder, and judges can be replaced without retraining.
Key points
- MOTHRAG handles multi-step question answering without .
- It runs through pay-per-call APIs instead of local GPUs or proprietary licenses.
- The reported benchmark average is 68.3 F1, close to GPU-heavy systems.
- The stated is $0.032 per query, or $0.018 per query on an economy tier.
- Proof trees make each answer step easier to inspect.