# Exact Is Easier: Credit Assignment for Cooperative LLM Agents

arXiv: https://arxiv.org/abs/2603.06859 · HTML: https://mlsanity.com/paper/2603.06859/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning for LLM agents, one of 16 directions in LLM agents and tool use.
- Builds on 31 of its 54 references that are in the map.
- First posted on arXiv 6 Mar 2026.

## Placed in

- LLM agents and tool use / [Reinforcement learning for LLM agents](https://mlsanity.com/agents/reinforcement-learning-for-llm-agents/)

## Builds on (most cited first, 30 of 31)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://mlsanity.com/paper/2308.08155/) (2023)
- [CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society](https://mlsanity.com/paper/2303.17760/) (2023)
- [Solving Quantitative Reasoning Problems with Language Models](https://mlsanity.com/paper/2206.14858/) (2022)
- [The Surprising Effectiveness of PPO in Cooperative, Multi-Agent Games](https://mlsanity.com/paper/2103.01955/) (2021)
- [ChatDev: Communicative Agents for Software Development](https://mlsanity.com/paper/2307.07924/) (2023)
- [Counterfactual Multi-Agent Policy Gradients](https://mlsanity.com/paper/1705.08926/) (2017)
- [Value-Decomposition Networks For Cooperative Multi-Agent Learning](https://mlsanity.com/paper/1706.05296/) (2017)
- [Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2003.08839/) (2020)
- [QPLEX: Duplex Dueling Multi-Agent Q-Learning](https://mlsanity.com/paper/2008.01062/) (2020)
- [The Complexity of Decentralized Control of Markov Decision Processes](https://mlsanity.com/paper/1301.3836/) (2013)
- [VinePPO: Refining Credit Assignment in RL Training of LLMs](https://mlsanity.com/paper/2410.01679/) (2024)
- [CMATH: Can Your Language Model Pass Chinese Elementary School Math Test?](https://mlsanity.com/paper/2306.16636/) (2023)
- [Reinforcing Multi-Turn Reasoning in LLM Agents via Fine-Grained Reward Structure and Credit Assignment](https://mlsanity.com/paper/2505.11821/) (2025)
- [Tree Search for LLM Agent Reinforcement Learning](https://mlsanity.com/paper/2509.21240/) (2025)
- [MALT: Improving Reasoning with Multi-Agent LLM Training](https://mlsanity.com/paper/2412.01928/) (2024)
- [MARFT: Multi-Agent Reinforcement Fine-Tuning](https://mlsanity.com/paper/2504.16129/) (2025)
- [LLM Collaboration With Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2508.04652/) (2025)
- [mmBERT: A Modern Multilingual Encoder with Annealed Language Learning](https://mlsanity.com/paper/2509.06888/) (2025)
- [Coevolving with the Other You: Fine-Tuning LLM with Sequential Cooperative Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2410.06101/) (2024)
- [Let It Flow: Agentic Crafting on Rock and Roll, Building the ROME Model within an Open Agentic Learning Ecosystem](https://mlsanity.com/paper/2512.24873/) (2025)
- [Stronger-MAS: Multi-Agent Reinforcement Learning for Collaborative LLMs](https://mlsanity.com/paper/2510.11062/) (2025)
- [Dr. MAS: Stable Reinforcement Learning for Multi-Agent LLM Systems](https://mlsanity.com/paper/2602.08847/) (2026)
- [Verified Critical Step Optimization for LLM Agents](https://mlsanity.com/paper/2602.03412/) (2026)
- [Unlocking the Power of Multi-Agent LLM for Reasoning: From Lazy Agents to Deliberation](https://mlsanity.com/paper/2511.02303/) (2025)
- [Who Deserves the Reward? SHARP: Shapley Credit-based Optimization for Multi-Agent System](https://mlsanity.com/paper/2602.08335/) (2026)

## Built on by (most cited first, 9 of 9)

- [From Reasoning to Agentic: Credit Assignment in Reinforcement Learning for Large Language Models](https://mlsanity.com/paper/2604.09459/) (2026)
- [Credit Without Ground Truth: Auditing Step-Level Credit Assignment in LLM Agents Against Executed Replay](https://mlsanity.com/paper/2608.19760/) (2026)
- [Reinforcement Learning for LLM-based Multi-Agent Systems through Orchestration Traces](https://mlsanity.com/paper/2605.02801/) (2026)
- [How Much, Then Where: Credit-Conserving Action-to-Token Allocation for Multi-Turn Agent Reinforcement Learning](https://mlsanity.com/paper/2608.07118/) (2026)
- [Cost-Utility Alignment in LLM Agent Trajectories:Profiling,Attribution,Diagnosis,Adaptation,and Evaluation](https://mlsanity.com/paper/2608.26195/) (2026)
- [Re-feeding Is Not Replaying: Measuring Replay Noise in Counterfactual Token-Credit Estimation](https://mlsanity.com/paper/2606.15621/) (2026)
- [Wrong but Useful: Trajectory Value Beyond Answer Correctness in Multi-Agent Messages](https://mlsanity.com/paper/2608.14375/) (2026)
- [Marginal-Contribution Policy Gradients under Filtered Feedback for Multi-Agent LLMs](https://mlsanity.com/paper/2604.22785/) (2026)
- [Learning What to Skip: Counterfactual Credit Assignment for Efficient Multi-Agent LLM Workflows](https://mlsanity.com/paper/2609.30734/) (2026)

## Abstract (arXiv)

Credit assignment for a team of LLM agents, what each message was worth, has mostly been treated as prediction: a learned critic, a trajectory-level score, or an agent ablation stands in for the counterfactual that defines credit, which is rarely run. Teams that communicate through a shared context are different: when everything a downstream agent reads is written into the trace, the trace is the state, and that counterfactual can be executed. A credit signal can then be judged as any estimator is: by bias, variance, and agreement with an independent reference. C3, credit assignment by counterfactual continuation, substitutes one message at a decision point and continues the run to the terminal reward, so its credit is unbiased, exact up to Monte Carlo error. Given the sampled alternatives, that error's variance follows a derived law with no term for the number of agents, and the observed noise follows the law on 6 workflows of 2 to 10 decision points. On the 2-agent chain, from 4 continuations, C3 ranks alternatives at 0.69 rank correlation against a 16-continuation reference, near the 0.73 at which 2 such references agree; a critic trained on the same rollouts reaches 0.29. Used as the advantage in training a 2-agent team, C3 beats MAGRPO, the stronger baseline in our comparison, and spends 37% fewer training tokens than MAPPO, since only the messages downstream of a decision point are regenerated. When the trace is the state, credit need not be predicted; it can be exact.