vishinvents/distil-qwen3-1.7b-posthog-prioritizer

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 17, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The vishinvents/distil-qwen3-1.7b-posthog-prioritizer is a 1.7 billion parameter specialist model, based on the Qwen3 architecture, developed by Distil Labs. It is specifically designed to rank product findings by impact, frequency, and effort, returning strict JSON output with a reason for each item. This model excels at ensuring exact coverage of input IDs in its ranked output, making it suitable for critical prioritization tasks where no item can be dropped or duplicated. It was fine-tuned using a teacher model (openai.gpt-oss-120b) and a combination of hand-authored and synthetically expanded data.

Loading preview...

Model Overview

The vishinvents/distil-qwen3-1.7b-posthog-prioritizer is a specialized 1.7 billion parameter model built on the Qwen3 architecture, developed by Distil Labs. Its primary function is to prioritize a batch of deduplicated product findings, such as bugs or gaps, based on their impact, frequency, and effort. The model is engineered to output a strict JSON format, providing a rank and a concise reason for each item.

Key Capabilities

  • Strict JSON Output: Guarantees valid JSON output where every input ID appears exactly once in the ranked list, preventing dropped, invented, or duplicated entries.
  • Contextual Prioritization: Ranks items based on predefined judgments, including prioritizing security issues, revenue blockers, and silent data corruption, while considering severity and reach.
  • Robustness: A 1.7B variant was chosen over a smaller 0.6B model due to its ability to maintain exact coverage even with adversarial batches containing near-duplicate findings.
  • Fine-tuned Performance: Trained using a combination of 22 hand-authored seed batches and 10,004 synthetically generated examples, with openai.gpt-oss-120b as the teacher model.

When to Use This Model

This model is ideal for use cases requiring automated, consistent, and explainable prioritization of product findings. It is particularly suited for:

  • Product Management: Automatically ranking bugs, feature requests, or other findings for development teams.
  • Incident Response: Prioritizing issues based on their severity and potential impact.
  • Automated Workflows: Integrating into systems that need to process and prioritize a stream of structured data, ensuring no item is overlooked.

Finetunes
1 models