jevIMAGE
Jev 跑 LLM-Wikirace 基准：450 局不到 1 美元
有人把 Jev 放进 LLM-Wikirace 跑了 450 局、8 小时，总花费不到 1 美元，结论是快且极便宜，但世界知识和规划能力不如前沿 LLM。
Ilija Bogunovic
@ilijabogunovic2026.09.22收录原推发布于 2026.09.22
Ilija Bogunovic @ilijabogunovic原推发布于 2026.09.22
@TypeSafeAI just released Jev, a new System 1 model, with Wikirace as a headline demo. So we ran it on ours: LLM-Wikirace benchmark. 450 games. 8 hours of wall-clock time. Under $1 total. Verdict: fast and absurdly cheap, but short on the world knowledge and planning that frontier LLMs bring. 🧵
x.com/ilijabogunovic/status/2102075332014624819采集于 2026.09.22 10:14