Not sure what your use case is, but I assume it is relatively simple. For agentic, long-context tasks the benchmarks tell us that American open models are not even in the ballpark. Consider the Artificial Analysis scores against the closest Chinese models (non-reasoning, similar sizes):
![Spanberger quote from the FT: “If we are classifying Tuesday as a success from a congressional standpoint, we will get [expletive] torn apart,” Spanberger said at the time. “Excuse the profanity, but that is the reality.”
“We need to not ever use the word socialist or socialism ever again,” she added. “Because while people think it doesn’t matter, it does matter.”](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:hrahz2kvjzbb2vy3wgzprbsx/bafkreib4ch7q6sh6xwlv3dpweohk55itdomsfukbtr322gsmjozty6t5xu)
![Article clip that reads: House Minority Leader Hakeem Jeffries (D-N.Y.) on Thursday said House Democrats will demand a ban on the deportation of American citizens as a condition of winning their support for legislation funding the Department of Homeland Security (DHS).
"In what country are we living in if it's controversial to prohibit [the deportation] of American citizens?" he said during a press
briefing in the Capitol. "That shouldn't even be a discussion."](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:hrahz2kvjzbb2vy3wgzprbsx/bafkreidx4swjhvqho6jqvzdmtcowu7runrg7wzwldakaahb73gshbl5u7q)
![Screenshot from truth social ai chat: A reasonable person, reviewing public behaviors like Trump's repeated Nobel Prize pursuits (e.g., nominations, medal acceptance offers, public coveting despite rejections), alongside consistent media observations of traits such as grandiose self-importance, excessive admiration-seeking, and entitlement (as cataloged by Kimmel matching all 9 DSM criteria; Frum's "clinical narcissism" analysis), would conclude these exemplify narcissistic tendencies.[1][2][3][5]
Implicit connection across the "3" (NPD traits, presidency, Nobel episode): Strong pattern of self-aggrandizement, where achievements are exaggerated for validation, per expert commentary (e.g., McAdams on constant self-focus).[3] This is behavioral inference from observable actions, not diagnosis—politically charged but widely noted in sources.[1][2][3][4]](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:hrahz2kvjzbb2vy3wgzprbsx/bafkreianzpbrqnn4mhswjs4yyxqiay5xikkx6e7734yfwpfi4fxi3yfnge)
