Contrastive Language Model 发布：8B 参数，推理速度达 Jev 的 9 倍
做自定义 harness 的可以看看 CLM-8B，比 Jev 快 9 倍，DeepSWE 跑到 81.6%，代码和模型都开源了。
Jacky Kwok 团队推出 Contrastive Language Model（CLM），一个用对比学习目标训练的 System One 模型。CLM-8B 在互联网规模数据上预训练，推理速度比用 RLCD 训练的 Jev 快 9 倍，在 DeepSWE 上取得 81.6%、Terminal-Bench 2.1 上取得 87.6% 的成绩。CLM 通过嵌入状态和候选动作并比较相似度来排序选择，而 Jev 接收情境加预定义问题后返回带概率的类型化决策。作者还将状态与动作的嵌入分离缓存，在动作集固定、状态持续演化的场景下显著降低推理延迟。团队还给出了 CLM 的缩放律：测试对比损失随训练算力、模型规模和数据量按幂律下降。
Pay attention to this new wave of System One models if you are building custom harnesses. First Jev. Now, Contrastive Language Model (CLM). CLM is 9x faster than Jev. CLM seems to be a better verifier than Jev, particularly at long-horizon tasks. How do Jev and CLM differ? CLM is contrastive, and Jev is trained with Reinforcement Learning for Calibrated Decisions (RLCD). Jev receives a situation plus predefined questions, and returns typed decisions with probabilities. CLM embeds the situation and candidate actions, compares their similarity, then ranks or selects the best match. The point is that there are several ways to attack this problem, which is exciting. You can see my recent guide on combining System One and System Two models for building custom harnesses. academy.dair.ai/resources/jev-… Jacky Kwok @jackyk02 Introducing Contrastive Language Model (CLM): an ultra-fast System One Model trained with a contrastive learning objective that connects states and actions. CLM-8B is pre-trained on internet-scale data and delivers up to 9× faster inference than Jev ⚡ while achieving comparable performance across computer-use, gaming, and tool-calling tasks. With lightweight fine-tuning, CLM-8B sets a new SOTA on challenging agentic coding benchmarks, such as DeepSWE (81.6%) and Terminal-Bench 2.1 (87.6%). In contrast, Jev fails to serve as an effective verifier for these long-horizon tasks. We also build an efficient training and serving infra for CLMs by disaggregating states and actions, allowing their embeddings to be cached and reused independently. This substantially reduces inference latency in settings where the state evolves continuously while the action set remains fixed. Finally, we establish scaling laws for CLMs and show that the test contrastive loss decreases predictably as a power law in training compute, model size, and dataset size. 📄 Blog: contrastive-lm.notion.site x 💻 Code github.com/Contrastive-LM… 8I 🗣️ Discor discord.gg/5dAQEDJBs o3J 🤗 Data & Mode huggingface.co/Contrastive-LM O3hu More details on CLM’s architecture, data recipe, and scaling laws in the thread below 🧵 Your browser does not support the video tag. 🔗 View on Twitter 🔗 View Quoted Tweet 💬 5 🔄 3 ❤️ 17 👀 2877 📊 8 ⚡
