Paper reports coding gains from ledger-based self-orchestration
A paper introduces ledger-based zero-shot self-orchestration, a training-free method in which fresh instances of the same model decompose problems and coordinate through a shared filesystem containing a plan, notes, and the current solution. The authors evaluated 11 open- and closed-weight models on the 100 latest hard LiveCodeBench problems and reported gains of up to 25.6 percentage points on pinned backends. They report that orchestrated Qwen3.8 Flash Next reached 93.0% pass@1 versus Fable 5's 90.4% at 9% of the cost, while locally served Qwen3.8-27B rose from 66.8% to 92.4%. The gains were not universal: some models were unchanged or performed worse.
The paper presents a training-free way to organize multiple instances of one model at inference time and reports substantial LiveCodeBench gains across 11 models. The authors also report that inference-time organization can approach frontier coding accuracy at a fraction of the cost or slightly exceed it on self-hostable weights.
GVS5H: Gorilla vs. 5 Humans
Results
Opus 5 ran on an earlier version of the scaffold, in a single pass served through OpenRouter rather than five on a pinned backend.
Abstract. Frontier coding performance is typically bought with larger proprietary models at high cost. We introduce ledger-based zero-shot self-orchestration, a training-free method in which fresh instances of one model decompose problems and coordinate through a shared filesystem holding a plan, notes and current solution. Across eleven open and closed-weight models on the 100 latest hard LiveCodeBench problems, the method yields gains of up to 25.6 percentage points on pinned backends and offers two routes to frontier-level accuracy. Orchestrated Qwen3.8 Flash Next reaches 93.0% pass@1 against Fable 5's 90.4% at 9% of the cost, and locally served, open-weight Qwen3.8-27B rises from 66.8% to 92.4%. Gains are not universal: some models are unchanged or worse. Transcript analysis attributes the gain to decomposition and persistent context. Inference-time organization can approach frontier coding accuracy at a fraction of the cost, or slightly exceed it on self-hostable weights. — the paper
#GVS5H#LiveCodeBench#Qwen3.8 Flash Next#Qwen3.8-27B#Fable 5#Opus 5#OpenRouter