[RELEASE] Supra-Router-51M - a tiny prompt routing model/orchestrator
Summary
SupraLabs releases Supra-Router-51M, a tiny 51M parameter model for routing prompts to appropriate larger or smaller models, enabling low-latency orchestration. A companion dataset is also released.
Similar Articles
SupraLabs/Supra-Router-51M
SupraLabs releases Supra-Router-51M, a 51.7M parameter micro-LLM for multi-task infrastructure routing, designed to decide whether to process prompts locally on edge or send them to cloud-hosted models. Fine-tuned on a small dataset, it uses multi-task sequence generation for robust routing.
@HarshalsinghCN: introducing tinyrouter i reverse engineered the routing architecture behind Skana AI's Fugu and built replication for o…
TinyRouter is a tiny 10K-parameter LLM router that learns to route each question to the best specialist model from a pool of open-source LLMs, using evolutionary training. It achieves performance matching or exceeding individual models on MMLU and math benchmarks.
[MASSIVE TINY RELEASE] - Supra2-Medium-Base - a tiny 25M parameters model competing heavily with our previous 50M model!
Supra2-Medium-Base is a new 25M parameter AI model based on qwen3 architecture, trained from scratch, that benchmarks competitively against a larger 50M parameter model.
@LiorOnAI: A ~10K parameter router can beat every individual open model on MMLU by learning which model should answer which questi…
A tiny ~10K parameter router called tinyrouter learns which open model to use per question on MMLU, outperforming individual models by optimizing allocation.
[NEW] Supra-50M Released!
SupraLabs released Supra-50M, a compact 50M-parameter causal language model with base and instruct versions, trained on 20B tokens from fineweb-edu, achieving competitive benchmarks against larger models like GPT-2 and SmolLM.