Hy3

User sentiment on HackerNews is sharply divided. The most upvoted positive comment notes Hy3 is 'pretty great, better than gpt-5.4-mini,' close enough to Sonnet 5 that 'I didn't notice much of a gap,' and 'exceedingly cheap so there is no reason not to use it, if you need a foss model.' However, multiple highly-upvoted comments are critical: one user reports it 'makes idiotic decisions all the time' when lacking task knowledge, another found it a 'mediocre performer in my benchmarks of security auditing,' outperformed by Gemma 4 31B and Qwen 3.6 27B. A commenter tracking OpenRouter rankings notes Hy3 'was topping the OpenRouter rankings' a month prior but 'has fallen to 8/9th,' concluding 'I don't see a reason where you would use this model over competitors.' Price economics are described as 'confusing,' with effective input price now matching DeepSeek Flash V4. The post-preview release is acknowledged as 'significantly better' with higher benchmark scores on DeepSWE, though one user cautions 'benchmarks are mostly meaningless — the only real benchmark is the actual work you give it.' Quantization resilience is flagged as an open question compared to DeepSeek V4's native FP4 architecture.

No aggregate ratings were found for this product during the last harvest.