- dataset: id: harborframework/terminal-bench-2.1 task_id: terminalbench_2_1 value: 80.9 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" - dataset: id: harborframework/terminal-bench task_id: terminalbench_4 value: 23.8 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" notes: "Run with mini-swe-agent v2.4.6, not Terminus." - dataset: id: datacurve/deep-swe task_id: deep_swe value: 65.5 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" notes: "Run with mini-swe-agent v2.4.6." - dataset: id: ScaleAI/SWE-bench_Pro task_id: SWE_Bench_Pro value: 60.3 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" - dataset: id: benchflow/skillsbench task_id: skillsbench_v1_1 value: 61.1 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" notes: "With-skills pass rate as reported on the model card, not the with/without-skills lift." - dataset: id: claw-eval/Claw-Eval task_id: general value: 78.3 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" - dataset: id: hkust-nlp/Toolathlon task_id: toolathlon_verified value: 70.4 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" notes: "No max_tokens limit per request." - dataset: id: collinear-ai/yc-bench task_id: medium value: 1025437 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" notes: "Final funds in USD; only the last 10 user turns kept as context." - dataset: id: Idavidrein/gpqa task_id: diamond value: 89.5 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card" - dataset: id: MathArena/aime_2026 task_id: MathArena/aime_2026 value: 95.2 date: "2026-10-25" source: url: https://huggingface.co/Nanbeige/Nanbeige4.3-33B name: "Nanbeige4.3-33B model card"