Fireworks Lab 发布了 Gen-1 Slides，这是一个用于幻灯片生成的开源模型，其性能与 Claude Opus 5 相当，但成本显著降低。该模型与 Genspark AI 共同开发，后者提供了产品判断和实际生产环境用于训练。开发过程涉及长时序强化学习，Fireworks Lab 负责基础设施和算法的共同开发，而 Genspark AI 则专注于产品方向和评估。 AI
影响 为成本效益高的幻灯片生成设定了新的基准，可能影响依赖演示文稿创建的工具和服务。
排序理由 前沿实验室模型发布，附带系统卡。
在 X — Fireworks (inference infra) 阅读 →
AI 生成摘要 · Google Gemini · 来自 5 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
The durable advantage is Genspark’s definition of a good deck: it lives in their evaluation, so they can retrain against it each time a stronger base model lands. Own the intelligence at the core of your product and the quality ceiling is yours to raise. Full write-up:
The hardest bugs were numerical. Long-horizon RL drifts when the engine generating rollouts and the one scoring them stop assigning the same probabilities to the same tokens. Aligning tokenization across both kept the training signal trustworthy.
Fireworks Lab co-developed the RL algorithm with Genspark and ran the training on frontier-grade infrastructure. Our embedded researchers engineered the reward, ran 100+ experiments, and read trajectories to catch the model gaming the score.
Genspark owned the product judgment. They brought their real production environment into training and shaped what the model optimized for: the standard for a good deck, the design principles behind it, and how to sharpen that standard over time.
Introducing Gen-1 Slides: an open model that matches Claude Opus 5 on slide generation, at ~1/17 of its input-token price. @genspark_ai post-trained it from a @MiniMax_AI M3 base with Fireworks Lab using long-horizon RL. Live today as the default in Genspark AI Slides. 🧵 https:/…