Qwopus3.8-27B-Flash

All projects

Apps & Tools

A fine-tuned Qwen3.8-27B optimized for fast, efficient agentic reasoning and tool-use.

byKyle Hessling

Playable

Qwopus3.8-27B-Flash media is blocked

Allow external media to connect to the provider and play this content.

Description

An optimized vision-language model fine-tuned using NeMo-RL and GSPO to reduce reasoning latency and token costs for agentic workloads. It features a 12.8% throughput gain and improved speculative decoding acceptance rates, specifically validated through an agentic software-engineering battery.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.

Details

Why this is interesting: It represents a standout effort to optimize model reasoning specifically for the cost and latency requirements of…
Attributed to: @KyleHessling1
Built with: K2 7B 3.8 Flash Gemini 3.8 Flash Cyber Omni 1.1 Flash Ling 3.0 Flash Fin Gemini 3.7 Flash Qwen3.7 Flash Opus 5 3.6 Flash 3.5 Flash Cyber Step 3.7 Flash Opus 4.8 Gemini 3.5 Flash Opus 4.7 Opus 4.6 Opus 4.5 2.5 Flash Qwen3.8-27B
Links: huggingface.co
View tweet
Tags: #multimodal #fine-tuned #agent #reasoning

Discovered: X
First posted: X · Sep 4, 2026
Listed: Added 10d ago
Build: Strong
Models: Strong