Reflection AI 开源 Beam-1.0-501B-A23B，Agent Arena 开源榜第 12
Exciting news: Beam-1.0-501B-A23B by @Reflection_AI just landed #12 among open weight models in Agent Arena with -3.67% Net Improvement, and #22 among open in C
Reflection AI 发布开源模型 Beam-1.0-501B-A23B，采用 Apache 2.0 许可，在 Agent Arena 开源权重模型中排第 12，Net Improvement 为 -3.67%，在 Code Arena: WebDev 开源榜排第 22，得分 1438 分。在 Agent Arena 榜单上它是排名最高的美国开源模型，领先 Hy3、MiMo V2.5 Pro 和 MiniMax M3，但 Net Improvement 误差区间达 ±3.84%，统计上与上方两个 Qwen3.8 模型更接近。Code Arena: WebDev 开源前 25 名中有 24 个来自中国实验室，Beam 1.0 是唯一例外，也是该榜排名最高的美国开源模型，略高于 Kimi K2.5 Thinking 和 GLM-5。
推荐理由
AI 评分 58
Beam 1.0 在 Agent Arena 与 Code Arena 的排名和误差区间，可看出美国开源模型在中文模型主导榜单中的位置。
主题
查看原帖隆重介绍 Beam：一款高效智能体开源模型，总参数 501B，激活参数 23B。 - 前沿推理效率 - 在编程与智能体任务上推进西方开源前沿 - 从零开始端到端训练 完整权重本月发布。 了解更多关于 Beam 的信息：https://t.co/c3Qx2cpM8G
查看原文
Introducing Beam: a highly efficient agentic open model with 501B total parameters and 23B active. - Frontier reasoning efficiency - Advances the Western open frontier on coding & agentic tasks - Trained end-to-end from scratch Full weights release this month. Learn more about Beam: https://t.co/c3Qx2cpM8G
正文
Exciting news: Beam-1.0-501B-A23B by @Reflection_AI just landed #12 among open weight models in Agent Arena with -3.67% Net Improvement, and #22 among open in Code Arena: WebDev with 1438 pts. It’s also fully open under Apache 2.0.
On the Agent Arena leaderboard, Beam 1.0 is the top-ranked US model among open models, landing at #12. That puts it ahead of Hy3, MiMo V2.5 Pro and MiniMax M3. Its margin of error is wide (±3.84% on Net Improvement), so it’s statistically closer to the two Qwen3.8 models just above it than its rank suggests.
On Code Arena: WebDev, 24 of the top 25 open models come from Chinese labs. The one exception is Reflection’s Beam 1.0 (Apache 2.0). It ranks #22 among open models, which makes it the top US open model on the board, just ahead of Kimi K2.5 Thinking and GLM-5.
Congrats to the @Reflection_AI team on this solid open weights release!