Mozilla报告：开源模型距前沿仅4个月，但营收占比极低
Mozilla just published a 91-page report and it says open-weight AI is now only a…
这份报告揭示了开源模型“叫好不叫座”的商业困境与产能现状，数据详实，对判断行业格局极具参考价值。
Mozilla just published a 91-page report and it says open-weight AI is now only about 4 months behind the frontier.
Mozilla 刚刚发布了一份 91 页的报告，指出开放权重 AI 目前仅落后于前沿模型约 4 个月。
- 8 of OpenRouter's 10 most-used models by August token volume were open-weight, and 7 were Chinese-built, while DeepSeek became the first open model to lead the platform in weekly requests.
- However, the economics are almost upside down. Open models handled roughly 20% of measured OpenRouter usage but captured only about 4% of model-layer revenue in the cited 2025 window.
- 截至八月按 token 用量计，OpenRouter 使用量最高的 10 个模型中有 8 个是开放权重的，其中 7 个由中国团队构建；而 DeepSeek 成为首个在周请求量上领先该平台的开放模型。
- 然而，经济状况几乎完全倒置。开放模型处理了约 20% 的 OpenRouter 测量使用量，但在引用的 2025 年窗口期内仅占模型层收入的约 4%。
Mozilla attributes much of that mismatch to pricing, with closed models costing roughly 6x more per call at about 90% capability parity. The comparison shows why high usage does not necessarily translate into high revenue when one class of models is dramatically cheaper. Mozilla also warns that the revenue measurement is older than its 2026 usage data, so the current revenue split may be different.
Mozilla 将这种不匹配很大程度上归因于定价问题，封闭模型每次调用的成本约为开放模型的 6 倍，尽管两者能力已大致持平（约 90%）。 该对比表明，当某一类模型价格显著更低时，高使用量并不必然转化为高收入。 Mozilla 还警告称，其收入测量数据早于 2026 年的使用量数据，因此当前的收入分配比例可能有所不同。
- Open models are spreading faster than they reach production: 79% of surveyed developers use them, but only 51% of open-model deployments reach production versus 63% for closed models.
- DeepSeek showed that a new pretraining run may no longer be necessary for a major capability jump, gaining roughly 10 index points and then another 8 through post-training passes.
- 开放模型的传播速度快于其进入生产环境的速度：79% 的受访开发者正在使用它们，但仅有 51% 的开放模型部署能进入生产环境，而封闭模型这一比例为 63%。
- DeepSeek 表明，实现重大能力跃升可能不再需要全新的预训练运行，通过后续训练阶段即可分别获得约 10 个指数点和另外 8 个指数点的提升。
-Even model diversification may provide less protection than assumed: Kimi K3 and Claude Fable 5 had a 0.72 per-task failure correlation, meaning supposed backup models often fail on the same problems.
-即使模型多样化提供的保护也可能低于预期：Kimi K3 和 Claude Fable 5 的任务级失败相关系数为 0.72，这意味着所谓的备用模型往往会在相同问题上失效。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力