开源模型离线12小时打榜ARC-AGI-2：数据与推理优化实战
Rules, Traces and Routing Statistics: Training and Calibration Data for Open Models on ARC-AGI-2
三个 Qwen 开源模型怎么在 12 小时离线预算里刷 ARC-AGI-2，规则数据、自推理轨迹、专家裁剪哪种真涨分，结果都在这了。
研究在四张 L4 GPU 上离线运行至多 12 小时，比较训练与校准数据对 ARC-AGI-2 成绩的影响，共报告 189 次半私有集运行。NVARC Qwen3-4B 用 12,000 条可执行规则和 95,993 个任务做语料，仅在 1,920 条真实任务记录上继续微调，把半私有分从 28.96 提到 33.06，合成任务全部无效。Qwen3.8-27B 用 1,036 条自身正确推理轨迹微调，加上强制思考帽和 Python 程序校验的推理流程，分数从 26.81 升至 44.03。125B 参数的 Qwen3.8-Flash-Next 通过按路由统计裁剪到 296 个专家加 FP8 索引缓存，把 KV 缓存扩到 1.96M tokens，最终拿下 56.67 的最好成绩。
Rules, Traces and Routing Statistics: Training and Calibration Data for Open Models on ARC-AGI-2
We study which training and calibration data help three open models solve ARC-AGI-2 tasks when the solver runs offline on four L4 GPUs for at most 12 hours. We compare data offline on public tasks and report 189 runs on the 120 hidden semi-private tasks. For the NVARC Qwen3-4B grid model, we built a corpus of 12,000 executable rules in a typed language and 95,993 tasks; a rule is admitted only if its behaviour on a fixed probe set differs from that of every admitted rule, and a task only if its pairs refute a shortlist of nearest competing rules and defeat 12 shortcut solvers. Continued fine-tuning on 1,920 real task records with revised four-GPU scheduling raised the semi-private score from 28.96 (mean of the released model) to 33.06 (mean of five runs, best 34.86); no checkpoint trained with synthetic tasks, even all 95,993 of them, improved on the real-task model. For the Qwen3.8-27B reasoning model, we fine-tuned on 1,036 of its own reasoning traces with correct final answers and compared sets that add traces of other models and synthetic tasks. With an inference procedure built around the model (a thinking cap enforced in the serving engine, a Python program checked on the demonstrations after each answer, a long first attempt per input), the score rose from 26.81 to 44.03, the level the ARC Prize Foundation measured for the base model without a time limit (42.36); most of the gain came from the inference procedure. For the Qwen3.8-Flash-Next mixture-of-experts model (125B parameters, 6B active), GPU memory limited concurrency. Pruning from 512 to 384 experts per layer by router-weighted activation saliency grew the key-value cache from 0.55M to 1.39M tokens; a 296-expert list adapted with ARC routing statistics, keeping experts for program tokens, plus an FP8 indexer cache, reached 1.96M. With a tool that checks programs, the model scored 56.67, our best result.
