User surprised by DeepSeek-V4-Flash-Vision-Exp performance when combined with DeepSeek harness

MaziyarPanahi · x · 2026-08-23

A developer reported switching from OpenAI's Codex due to quota limits to a combination of DeepSeek-V4-Flash-Vision-Exp and the DeepSeek harness. The user expressed shock at the results, suggesting the new model setup delivers impressive performance, potentially for coding or vision tasks.

Related event: DeepSeek V4 Flash Vision Impresses in Hands-On Tests(4 posts)→

Original post →

More from Models

Models channel →