X · @teortaxesTex
· X / Twitter
insane actually how much Anthropic has accelerated in 2026 open source models are so often within striking distance of GPT 5.5, but Opus 4.8 is consis…
insane actually how much Anthropic has accelerated in 2026open source models are so often within striking distance of GPT 5.5, but Opus 4.8 is consistently an aspirational target (on coding at least)Lentils: GLM-5.2 and Kimi K2.7 Code get good scores on FrontierCode Extended"FrontierCode is a benchmark that measures ho