# TACO: Tool-Augmented Credit Optimization for Agentic Tool Use

arXiv: https://arxiv.org/abs/2606.30251 · HTML: https://mlsanity.com/paper/2606.30251/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal and embodied agents, one of 16 directions in LLM agents and tool use.
- Builds on 36 of its 51 references that are in the map.
- First posted on arXiv 29 Jun 2026.

## Placed in

- LLM agents and tool use / [Multimodal and embodied agents](https://mlsanity.com/agents/multimodal-and-embodied-agents/)

## 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)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models](https://mlsanity.com/paper/2504.10479/) (2025)
- [MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts](https://mlsanity.com/paper/2310.02255/) (2023)
- [ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning](https://mlsanity.com/paper/2203.10244/) (2022)
- [Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://mlsanity.com/paper/2501.12599/) (2025)
- [MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?](https://mlsanity.com/paper/2403.14624/) (2024)
- [BLINK: Multimodal Large Language Models Can See but Not Perceive](https://mlsanity.com/paper/2404.12390/) (2024)
- [Measuring Multimodal Mathematical Reasoning with MATH-Vision Dataset](https://mlsanity.com/paper/2402.14804/) (2024)
- [Kimi K2.5: Visual Agentic Intelligence](https://mlsanity.com/paper/2602.02276/) (2026)
- [V*: Guided Visual Search as a Core Mechanism in Multimodal LLMs](https://mlsanity.com/paper/2312.14135/) (2023)
- [VLMEvalKit: An Open-Source Toolkit for Evaluating Large Multi-Modality Models](https://mlsanity.com/paper/2407.11691/) (2024)
- [MME-RealWorld: Could Your Multimodal LLM Challenge High-Resolution Real-World Scenarios that are Difficult for Humans?](https://mlsanity.com/paper/2408.13257/) (2024)
- [We-Math: Does Your Large Multimodal Model Achieve Human-like Mathematical Reasoning?](https://mlsanity.com/paper/2407.01284/) (2024)
- [Pixel Reasoner: Incentivizing Pixel-Space Reasoning with Curiosity-Driven Reinforcement Learning](https://mlsanity.com/paper/2505.15966/) (2025)
- [LogicVista: Multimodal LLM Logical Reasoning Benchmark in Visual Contexts](https://mlsanity.com/paper/2407.04973/) (2024)
- [Mini-o3: Scaling Up Reasoning Patterns and Interaction Turns for Visual Search](https://mlsanity.com/paper/2509.07969/) (2025)
- [Reinforcing Multi-Turn Reasoning in LLM Agents via Fine-Grained Reward Structure and Credit Assignment](https://mlsanity.com/paper/2505.11821/) (2025)
- [MathCoder-VL: Bridging Vision and Code for Enhanced Multimodal Mathematical Reasoning](https://mlsanity.com/paper/2505.10557/) (2025)
- [Agent0-VL: Exploring Self-Evolving Agent for Tool-Integrated Vision-Language Reasoning](https://mlsanity.com/paper/2511.19900/) (2025)
- [TemplateRL: Structured Template-Guided Reinforcement Learning for LLM Reasoning](https://mlsanity.com/paper/2505.15692/) (2025)
- [CodeV: Code with Images for Faithful Visual Reasoning via Tool-Aware Policy Optimization](https://mlsanity.com/paper/2511.19661/) (2025)
- [AStar: Boosting Multimodal Reasoning with Automated Structured Thinking](https://mlsanity.com/paper/2502.02339/) (2025)
- [PyVision-RL: Forging Open Agentic Vision Models via RL](https://mlsanity.com/paper/2602.20739/) (2026)
- [Act Wisely: Cultivating Meta-Cognitive Tool Use in Agentic Multimodal Models](https://mlsanity.com/paper/2604.08545/) (2026)
- [RadialRouter: Structured Representation for Efficient and Robust Large Language Models Routing](https://mlsanity.com/paper/2506.03880/) (2025)
- [PACR: Progressively Ascending Confidence Reward for LLM Reasoning](https://mlsanity.com/paper/2510.22255/) (2025)
- [Step Potential Advantage Estimation: Harnessing Intermediate Confidence and Correctness for Efficient Mathematical Reasoning](https://mlsanity.com/paper/2601.03823/) (2026)
- [Exploring Knowledge Purification in Multi-Teacher Knowledge Distillation for LLMs](https://mlsanity.com/paper/2602.01064/) (2026)
- [PatchCue: Enhancing Vision-Language Model Reasoning with Patch-Based Visual Cues](https://mlsanity.com/paper/2603.05869/) (2026)

## Built on by (most cited first, 5 of 5)

- [The Illusion of Visual Tool-Use: A Causal Audit of Thinking with Images](https://mlsanity.com/paper/2608.06270/) (2026)
- [SkillGate: Training In-Policy Skill Selection in Long-Horizon Agents](https://mlsanity.com/paper/2608.18852/) (2026)
- [MLToolBench: Learning Tool-Augmented Agents for Machine Learning Development](https://mlsanity.com/paper/2609.36679/) (2026)
- [Decoupled Visual Processing: Efficient Multimodal Adaptation via Modality-Specific Transformer Substitution](https://mlsanity.com/paper/2607.26596/) (2026)
- [When Should a VLM Look? Paying Only for Visual Calls That Were Needed and Used](https://mlsanity.com/paper/2609.22910/) (2026)

## Abstract (arXiv)

Agentic multimodal models perform diverse operations on an image via code and reason over the returned view, an effective paradigm for fine-grained visual question answering. However, code operations can be useful, redundant, or misleading. Outcome-only rewards cannot precisely distinguish these cases, and existing process rewards either fail to attribute final correctness to individual tool calls, or require an external judge model. To address this, we introduce Tool-Augmented Credit Optimization (TACO), a GRPO variant for code-tool agents built on two coupled advantage channels. The first, Differential Answer-Probe Reward (DAPR), is a self-supervised, judge-free tool-contribution advantage that credits each tool call by its own effect on answering correctly. Probe tokens inserted into the model's reasoning elicit its predictions with and without the tool, and the difference in outcome reward is taken as the call's value: positive for a useful call, negative for a misleading one, and zero for one that changes nothing. This reuses the existing answer checker with no auxiliary judge, and, being a difference rather than an absolute probe score, is naturally robust to probe-hacking. The second is the outcome advantage from the final answer, distributed by Outcome-Gated Advantage Routing (OGAR): a parameter-free rule that, conditioned on the call's outcome, delivers this credit only to the responsible segments, suppressing wasted tool calls without any cost term. We train TACO through a two-stage SFT+RL pipeline. Extensive experiments across perception, reasoning, and general multimodal benchmarks show that it yields consistent accuracy gains and learns to invoke its tools only when they help.