# WHALE: A Simple Recipe for Joint Harness-Weight Optimization

arXiv: https://arxiv.org/abs/2609.00196 · HTML: https://mlsanity.com/paper/2609.00196/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 36 of its 43 references that are in the map.
- First posted on arXiv 31 Aug 2026.

## Builds on (most cited first, 30 of 36)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [DAPO: An Open-Source LLM Reinforcement Learning System at Scale](https://mlsanity.com/paper/2503.14476/) (2025)
- [Billion-scale similarity search with GPUs](https://mlsanity.com/paper/1702.08734/) (2017)
- [SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering](https://mlsanity.com/paper/2405.15793/) (2024)
- [Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning](https://mlsanity.com/paper/2503.09516/) (2025)
- [Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Steps](https://mlsanity.com/paper/2011.01060/) (2020)
- [When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories](https://mlsanity.com/paper/2212.10511/) (2022)
- [Measuring and Narrowing the Compositionality Gap in Language Models](https://mlsanity.com/paper/2210.03350/) (2022)
- [MuSiQue: Multihop Questions via Single-hop Question Composition](https://mlsanity.com/paper/2108.00573/) (2021)
- [Large Language Models as Optimizers](https://mlsanity.com/paper/2309.03409/) (2023)
- [DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines](https://mlsanity.com/paper/2310.03714/) (2023)
- [RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment](https://mlsanity.com/paper/2304.06767/) (2023)
- [GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning](https://mlsanity.com/paper/2507.19457/) (2025)
- [Reinforced Self-Training (ReST) for Language Modeling](https://mlsanity.com/paper/2308.08998/) (2023)
- [RRHF: Rank Responses to Align Language Models with Human Feedback without tears](https://mlsanity.com/paper/2304.05302/) (2023)
- [AFlow: Automating Agentic Workflow Generation](https://mlsanity.com/paper/2410.10762/) (2024)
- [Automated Design of Agentic Systems](https://mlsanity.com/paper/2408.08435/) (2024)
- [ReTool: Reinforcement Learning for Strategic Tool Use in LLMs](https://mlsanity.com/paper/2504.11536/) (2025)
- [Beyond Human Data: Scaling Self-Training for Problem-Solving with Language Models](https://mlsanity.com/paper/2312.06585/) (2023)
- [Meta-Harness: End-to-End Optimization of Model Harnesses](https://mlsanity.com/paper/2603.28052/) (2026)
- [Coordinate Descent Algorithms](https://mlsanity.com/paper/1502.04759/) (2015)
- [AgentSquare: Automatic LLM Agent Search in Modular Design Space](https://mlsanity.com/paper/2410.06153/) (2024)
- [Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses](https://mlsanity.com/paper/2604.25850/) (2026)
- [Language Agents as Optimizable Graphs](https://mlsanity.com/paper/2402.16823/) (2024)
- [Fine-Tuning and Prompt Optimization: Two Great Steps that Work Better Together](https://mlsanity.com/paper/2407.10930/) (2024)
- [Evolving Agents in the Dark: Retrospective Harness Optimization via Self-Preference](https://mlsanity.com/paper/2606.05922/) (2026)

## Built on by (most cited first, 11 of 11)

- [Recursive self-improvement of AI research agents](https://mlsanity.com/paper/2609.26457/) (2026)
- [Harness-Zero: Harness Distillation via Agent-as-Harness](https://mlsanity.com/paper/2609.24974/) (2026)
- [AgentVidBench: A Multi-Hop Video Question Answering Benchmark for Evaluating MLLM Agents](https://mlsanity.com/paper/2609.21386/) (2026)
- [PUBG Ally: A Conversational Embodied Agent as an AI Teammate](https://mlsanity.com/paper/2609.29837/) (2026)
- [RLHarness: Co-evolving Procedural Skills with Reinforcement Learning for Long-horizon Multimodal Reasoning](https://mlsanity.com/paper/2609.32326/) (2026)
- [Harness Learning Enables Generalizable Test-Time Adaptation](https://mlsanity.com/paper/2609.35738/) (2026)
- [VACE: Validation-Gated Alternating Co-Evolution of Agent Models and Harnesses](https://mlsanity.com/paper/2609.37105/) (2026)
- [ActiveSaddler: Automated Curriculum Learning for Agent Harness Optimization](https://mlsanity.com/paper/2610.00906/) (2026)
- [Not All Experience Belongs in the Weights: Component Routing for Self-Improving GUI Agents](https://mlsanity.com/paper/2610.01787/) (2026)
- [ASCENT: Online Test-Time Training of Long-Horizon Agents via Self-Distillation of Verified Experience](https://mlsanity.com/paper/2610.05303/) (2026)
- [Harness Evolution Hits a Ceiling: When Weight Training Should Begin](https://mlsanity.com/paper/2610.11655/) (2026)

## Abstract (arXiv)

Agent performance depends jointly on the model parameters and the executable harness code that manages context and control flow. Optimizing either component in isolation can leave the system bottlenecked by its frozen counterpart: weight updates can change which harness is effective, while harness updates can change which model capabilities are exposed. Existing joint-adaptation methods optimize weights and textual prompts but leave the broader harness fixed. We propose Weight-Harness Alternating LEarning (WHALE), a simple recipe that alternates two phases: updating the model under the current harness, then searching for a better harness under the updated model. We instantiate these two phases with online rejection-sampling fine-tuning and Meta-Harness, respectively. When to switch is a key design choice: to separate real improvements from noise without over-optimizing against a changing counterpart, WHALE uses either fixed phase durations or an adaptive patience rule over training signals. Using Qwen3.5-2B/4B agents across three domains (search question answering, mathematical reasoning, and chess puzzles), WHALE outperforms weight-only, harness-only, and Fast-Slow Training by 4.15-24.38 percentage points in best mean@8 accuracy. Either component can be the bottleneck: harness search matches peak weight-only accuracy with far fewer rollouts in SearchQA, but improves math accuracy only after a weight update. Small interleaved updates also outperform stagewise weight-then-harness optimization in accuracy and rollout cost. The code is available at https://github.com/krafton-ai/WHALE.