The tweet was deleted by the author.
But we saved everything 🙂.
Sonnet 4.5 has achieved a notable position by leading the LiveBench rankings in agentic coding.
The software not only excels in its category but also surpasses its competitors, with GPT-5 closely following. Additionally, Sonnet 4.5 (thinking) has gained the top spot in the non-agentic coding category, further solidifying its dominance in the field.
Sonnet 4.5’s ascendancy in agentic and non-agentic coding thus reflects a broader trend of innovation within artificial intelligence, reminiscent of breakthroughs observed in pioneering models such as Seadream 4. The recent results also underscore ongoing debates about the evolving prerequisites for success in tech, echoing discussions around nontraditional pathways to achievement within the industry.