AI News AI资讯 1d ago Updated 2h ago 更新于 2小时前 52

Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower Cost Cognition发布SWE-2:基于Kimi K3后训练的编码模型,在FrontierCode上与Fable 5.1持平,成本降低64%

Cognition released SWE-2, a coding model post-trained via RL from Moonshot AI's 2.8T-parameter Kimi K3, achieving 50.0% on FrontierCode 1.1 Main—within 1 point of Fable 5.1 at 64% lower cost SWE-2 is the first Cognition model with selectable reasoning-effort levels trained in a single RL run using Pareto-informed linear cost penalties, reducing turns by 58% and cost by 81% versus SWE-1.7 The model demonstrates significant RL headroom on top of the massive K3 base, adding 5–6 points on most bench Cognition发布SWE-2,基于2.8T参数Kimi K3经强化学习后训练,FrontierCode 1.1 Main达50.0%,接近Fable 5.1(50.9%)且成本降低64% 首款支持多推理努力级别(medium/high/max)的Cognition模型,通过单次RL训练同步优化成本-性能前沿 SWE-2 medium相比SWE-1.7在FrontierCode上任务轮次减少58%、成本降低81%,首轮编辑中位步骤从48降至18 技术核心包括Pareto感知成本惩罚、长度加权奖励基线、DSpark推测解码及NVFP4/FP8量化训练 模型未开源且无独立API,仅通过Devin

82
Hot 热度
68
Quality 质量
72
Impact 影响力

Analysis 深度分析

TL;DR

  • Cognition released SWE-2, a coding model post-trained via RL from Moonshot AI's 2.8T-parameter Kimi K3, achieving 50.0% on FrontierCode 1.1 Main—within 1 point of Fable 5.1 at 64% lower cost
  • SWE-2 is the first Cognition model with selectable reasoning-effort levels trained in a single RL run using Pareto-informed linear cost penalties, reducing turns by 58% and cost by 81% versus SWE-1.7
  • The model demonstrates significant RL headroom on top of the massive K3 base, adding 5–6 points on most benchmarks while showing focused exploration behavior with fewer detours
  • SWE-2 remains proprietary with no open weights or standalone API, accessible only through Devin Desktop, CLI, and rolling out to Web and Fusion
  • Despite strong performance on Terminal-Bench 2.1 (92.8%) and FrontierCode, SWE-2 shows a clear weakness on Terminal-Bench 4 (27.3%), trailing top competitors by ~30 points

Why It Matters

SWE-2 demonstrates that scaling reinforcement learning to multi-trillion-parameter models yields substantial performance gains, proving that even state-of-the-art bases like Kimi K3 have significant untapped potential through task-specific RL post-training. For AI practitioners, it highlights the emerging pattern of "effort-level" models that let users trade compute for performance, and the importance of specialized coding benchmarks in evaluating real-world agent capability beyond generic LLM scores.

Technical Details

  • Training pipeline: Post-trained from Kimi K3 (2.8T-parameter MoE) using a novel single-run RL algorithm that jointly trains three effort levels (medium/high/max) with slope-matched linear cost penalties, where the reward function is R = S − λC (binary success minus cost-weighted penalty)
  • Cost-performance frontier: Each effort level's λ is set to the local slope of the base model's Pareto curve, making the iso-reward line tangent to the frontier so reward can only improve by pushing the frontier upward—not just shifting along it
  • Efficiency gains: Length-weighted reward baseline (sum(R×L)/sum(L)) stabilizes training without extra compute; DSpark speculative decoding with a SpecForge-retrained draft model achieves 15% longer accept lengths; NVFP4 and FP8 kernels with quantization-aware training reduce train-inference mismatch
  • Benchmark performance: FrontierCode 1.1 Main 50.0%, DeepSWE 1.1 73.0%, Terminal-Bench 2.1 92.8%, Terminal-Bench 4 27.3%; leads on Terminal-Bench 2.1 and beats Kimi K3 base by 5–6 points across most metrics
  • Behavioral improvements: SWE-2 medium makes its first real edit after a median of 18 steps versus 48 for SWE-1.7; shows stronger end-to-end test coverage, resourcefulness when tools are blocked, and verification discipline (re-deriving rather than re-asserting conclusions)

Industry Insight

The "effort-level" paradigm—where a single model serves multiple reasoning budgets trained simultaneously—is likely to become a standard pattern for production AI agents, allowing teams to dynamically allocate compute based on task complexity rather than maintaining separate model variants. The proprietary-only distribution strategy (Devin exclusively) signals that top-tier coding models are increasingly treated as competitive moats rather than open commodities, reinforcing the trend where value shifts from raw model weights to the agent infrastructure and RL recipes built around them. Finally, the significant gap between Terminal-Bench 2.1 (92.8%) and Terminal-Bench 4 (27.3%) suggests that current benchmarks still fail to capture consistent long-horizon reasoning ability, prompting the industry to develop more rigorous multi-step evaluation suites.

TL;DR

  • Cognition发布SWE-2,基于2.8T参数Kimi K3经强化学习后训练,FrontierCode 1.1 Main达50.0%,接近Fable 5.1(50.9%)且成本降低64%
  • 首款支持多推理努力级别(medium/high/max)的Cognition模型,通过单次RL训练同步优化成本-性能前沿
  • SWE-2 medium相比SWE-1.7在FrontierCode上任务轮次减少58%、成本降低81%,首轮编辑中位步骤从48降至18
  • 技术核心包括Pareto感知成本惩罚、长度加权奖励基线、DSpark推测解码及NVFP4/FP8量化训练
  • 模型未开源且无独立API,仅通过Devin Desktop/CLI提供服务,付费用户至2026年10月10日免费使用

为什么值得看

SWE-2展示了多万亿参数模型经针对性强化学习后在编码任务上的显著突破,为行业提供了兼顾性能与成本效率的可行路径。其多推理努力级别统一训练框架及行为控制机制,对开发下一代Agentic AI系统具有重要参考价值。

技术解析

  • 基准测试表现:在FrontierCode 1.1 Main(50.0%)、DeepSWE 1.1(73.0%)、Terminal-Bench 2.1(92.8%)上领先或接近竞品;Terminal-Bench 4(27.3%)显著落后于Fable 5.1(55.8%)和GPT-6 Astra(57.9%)
  • RL训练机制:采用线性成本惩罚奖励函数R=S−λC(S为二进制成功信号,C混合推理成本与时间),λ设置为基座模型Pareto前沿局部斜率,使单一训练同步优化多个努力级别
  • 训练稳定性优化:使用长度加权组内基线sum(R×L)/sum(L)降低梯度方差与KL散度;Prefill Delayer批处理提升10-20%吞吐量;SpecForge重训练草稿模型以支持15%更长接受长度
  • 数据与可信度:RL环境数量增至三倍,叠加指令遵循覆盖;构建验证器反馈飞轮修正假阳性/假阴性;在145个涉华政治敏感问题测试中通过率98.0%(英文99.8%、简体中文95.2%、繁体中文99.1%)

行业启示

  • 封闭生态持续主导高端编码AI:SWE-2无权重开放策略表明,顶尖模型仍将通过垂直整合的Agent平台交付,开发者需关注平台接入而非本地部署
  • 成本效率成为核心竞争力:在保持与头部模型相近性能的同时实现64%成本下降,凸显强化学习优化推理路径的商业价值,行业竞争焦点正从纯精度转向性价比
  • 多努力级别统一训练是可行范式:单次RL运行同步调节不同成本预算下的模型行为,为资源受限场景提供自适应推理方案,可推广至其他专业领域模型开发

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

Code Generation 代码生成 Agent Agent Open Source 开源 Research 科学研究 Benchmark 基准测试