Inkling 2 Small Launch - PMPP-Hard - AUC (task-hours): leaderboard

69 GPU kernel-optimization tasks; four independent attempts; 10-hour budget. Each model uses its maximum available reasoning setting, PI harness and shared compaction policy. Task/grading fixes applied; two independent LLM audits per solution and third adjudication on disagreement. AUC spans 0–10 hours (max 690 task-hours); mean and pass@4 are counts out of 69. No maximum-effort name is converted into an unsupported provider-specific setting.

Metric: AUC (task-hours). Source: thinkingmachines.ai. Saturation forecast: Rough model projection: around 2026. 9 models tracked.

Top models

#ModelScoreOverall rank
1GPT-6 Astra (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)616.39
2Kimi K3.1 Pro (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)583.86
3Claude Opus 5.5 (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)499.65
4DeepSeek V4.1 Flash (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)497.36
5Inkling 2 Small (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)494.27
6Mistral Large 4 (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)478.39
7Kimi K3 (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)453.58
8GLM 5.3 (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)430.96
9GLM 5.2 (Benchmark Sweep Run: source=Inkling2SmallPMPP; PI harness; max available reasoning; 4 attempts; 10h)389.12

Interactive version: theaggregate.ai/benchmark?slug=inkling-2-small-launch-pmpp-hard-auc-task-hours · How It Works · Data refreshed daily, snapshot 2026-11-06.