{ "id": "qwen3-4b-file-ops-distilled", "updated": "2026-06-19", "artifact": "hf://models/sarthakagrawal927/qwen3-4b-file-ops-distilled", "base": "Qwen/Qwen3-4B-Instruct-2507", "precision": "bf16", "training_method": "frontier/gold trajectory distillation on GorillaFileSystem multi-turn tasks", "scores": [ { "suite": "file_ops_hard_gate", "n": 12, "stock_4b": 0.58, "distilled_4b": 1.0, "frontier": 1.0, "source": "docs/learn/tool-calling-frontier-parity.md#81-climbing-the-cliff--frontier-trajectory-distillation-2026-06-16" }, { "suite": "file_ops_hardgen_heldout", "n": 40, "stock_4b": 0.6, "distilled_4b": 0.95, "source": "docs/learn/tool-calling-frontier-parity.md#81-climbing-the-cliff--frontier-trajectory-distillation-2026-06-16" }, { "suite": "out_of_domain_breadth", "n": 52, "stock_4b": 0.596, "distilled_4b": 0.423, "delta": -0.173, "source": "docs/learn/tool-calling-frontier-parity.md#84-breadth--narrow-distillation-causes-negative-transfer-2026-06-16" } ], "verdict": "ship only as a routed file-ops specialist; do not use as the general planner", "caveats": [ "The out-of-domain breadth ceiling is not frontier-validated yet.", "The negative-transfer result is relative and directly comparable: same 52 tasks, same prompt, stock vs distilled.", "The model is a fused full HF/MLX directory, not a small adapter package." ] }