Davinci-Tool: Scaling Agentic Tool-Use Supervision with Verifiable Executable Environments

Xiaoyang Yuan; Ji Zhao; Xingyu Hu; Wenhan Wang; Yutao Hu; Qiaosheng Zhang; Pengfei Liu; Yu Qiao

cs.LGOct 12, 2026

Davinci-Tool: Scaling Agentic Tool-Use Supervision with Verifiable Executable Environments

Authors: Xiaoyang Yuan, Ji Zhao, Xingyu Hu, Wenhan Wang, Yutao Hu, Qiaosheng Zhang, Pengfei Liu, Yu Qiao,

Organizations: Tongji University · Shanghai Innovation Institute · University of Science and Technology of China · Southeast University

Abstract

Reliable agentic tool use requires training data that pair diverse multi-step workflows with executable evidence that the requested outcomes were achieved. Scaling either trajectory coverage or verification alone is insufficient as new environments introduce distinct states, rules, and dependencies that must also be reflected in the success criteria. To address this challenge, we introduce Davinci-Tool, a 30B tool-use model, and Davinci-ToolData, an execution-grounded dataset designed to scale environment coverage and outcome verification together. Davinci-ToolData preserves restorable states, execution traces, outcomes, and verifiers as a unified substrate for supervision. Two complementary pipelines populate this substrate: Execution-Grounded Task Synthesis (EGTS) derives verified demonstrations from independently authored systems, and Executable Environment Synthesis (EES) synthesizes auditable environments, state-grounded tasks, and executable verifiers across single and coordinated domains. These pipelines yield 30,395 verified trajectories spanning 2,260 invoked tool names. We first train on these demonstrations with supervised fine-tuning, then reuse EES states and verifiers to compare alternative actions and construct outcome-grounded preferences for Direct Preference Optimization. Davinci-Tool improves Qwen3-Coder-30B-A3B-Instruct from 52.4 to 88.2 on $τ^2$ -bench, from 30.4 to 62.8 on BFCL v4 multi-turn, and from 16.9 to 56.2 on MCP-Atlas, outperforming Qwen3-Coder-480B-A35B-Instruct on all three benchmarks with only 1/16 of its total parameters. Code is available at https://github.com/MassLab-SII/Davinci-Tool.

Figures & tables

Figure 1: Performance on

\tau^{2}

-bench, BFCL v4 multi-turn, and MCP-Atlas. Davinci-Tool-30B after DPO is shown with hatching, with its SFT checkpoint included for comparison. Detailed results and score provenance are reported in Table 2 .

Figure 2: Construction of Davinci-ToolData and training of Davinci-Tool. EGTS converts explored outcomes in authored systems into verified task-conditioned replay demonstrations, while EES constructs audited executable environments for verified single-domain and cross-domain trajectories. The resulting demonstrations support SFT, and restorable EES states and verifiers are subsequently reused for decision-level DPO.

Source	Traj.	Tokens	Calls	Tools
EGTS	13,824	10,063.9	6.04	350
EES-Single	8,026	10,372.7	5.87	1,610
EES-Cross	8,545	20,444.7	7.01	1,683
Full SFT corpus	30,395	13,063.8	6.26	2,260

Table 1: Verified SFT corpus. Token and call counts are per-trajectory means. Tool counts represent distinct invoked serialized names.

Figure 3: Structure of the

246

retained EES-Cross compositions. Panels (a), (b), and (c) show the number of constituent domains per environment, the number of application categories per environment, and category pairs appearing in at least

5

compositions, respectively. Pair counts may overlap when an environment spans more than two categories. Category codes are defined in Table 5 .

Reference models
Model	Training	$\boldsymbol{\tau^{2}}$ -bench			BFCL v4 multi-turn					MCP-Atlas
		Retail	Telecom	Avg.	Base	Long-Context	Miss-Func	Miss-Param	Avg.	Avg.
Qwen3-Coder-480B-A35B-Instruct	—	77.5	39.5	58.5	55.5	54.5	32.0	32.0	43.5	28.1
Kimi-K2-Instruct	—	70.6	65.8	68.2	42.0	38.5	23.0	30.0	33.4	59.6
MiniMax-M2.1	—	81.6	87.0	84.3	76.5	69.0	71.0	69.0	71.4	59.6
GLM-4.7	—	79.8	98.2	89.0	69.5	67.0	66.0	59.0	65.4	55.1

Table 2: Main results and model transfer. Reference models are strong open- and closed-source models included for external comparison. Underlining marks the best reference-model score in each metric, while bold marks the best score within each of the Qwen3-8B, Qwen3-14B, and primary 30B training groups. Light and dark shading denote Davinci-ToolData SFT and SFT+DPO, respectively.

SFT data			Benchmarks
EGTS	EES-Single	EES-Cross	$\boldsymbol{\tau^{2}}$ -bench	BFCL v4	MCP-Atlas
$\checkmark$			78.9	57.4	28.1
	$\checkmark$		82.0	54.5	29.2
		$\checkmark$	83.4	46.6	47.2
	$\checkmark$	$\checkmark$	82.9	55.6	50.6
$\checkmark$	$\checkmark$		79.4	57.6	31.5

Table 3: SFT data-source mixtures. Bold marks the strongest evaluated mixture on each benchmark.

Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

Environment family	Representative interactions	Representative adapted sources
Personal and business services	Contacts, messages, reminders, calendars, banking, travel, health, and other structured service operations	ToolSandbox ( Lu et al., 2025 ) , API-Bank ( Li et al., 2023 )
API discovery and orchestration	API search, schema inspection, argument grounding, selection, and multi-API composition	StableToolBench ( Guo et al., 2024 )
Developer and data workspaces	Filesystem operations, repositories, relational data, and browser-backed developer workflows	MCPMark ( Wu et al., 2026b )
Terminal and professional production	Shell and text processing, system operations, reports, office files, PDFs, and data deliverables	Terminal-Bench ( Merrill et al., 2026 )
Web interaction and commerce	Page navigation, form interaction, product search, constraint matching, and shopping workflows	MiniWoB ( Shi et al., 2017 ) , WebShop ( Yao et al., 2022 )
Scientific simulation	Procedural experiments, hypothesis testing, embodied discovery, and stateful scientific interaction	ScienceWorld ( Wang et al., 2022 ) , DiscoveryWorld ( Jansen et al., 2024 )

Appendix

Table 4: Environment families represented in the EGTS library. Each source is adapted to the common execution and evidence-capture interface.

Code	Application category	$\boldsymbol{D_{c}}$	$\boldsymbol{S_{c}}$
HC	Healthcare & Clinical	30	115
IT	IT Infrastructure & Software Engineering	20	80
SOC	Social & Communication	14	56
ECOM	E-commerce, Retail & Marketplace	11	47
ENT	Enterprise, HR & Business Ops	12	47
PHW	Personal Health & Wellness	13	47

Appendix

Table 5: EES domain categories and composition usage.

D_{c}

is the number of domain definitions in category

c

, and

S_{c}

counts their occurrences across the 246 retained EES-Cross compositions. Rows follow decreasing

S_{c}

. Category codes are used in Figure 3 .

Metric	EES-Single	EES-Cross
Domain transitions (%)	0.00	14.50
Cross-domain entity reuse (%)	0.00	5.23
Trajectories with cross-domain transfer (%)	0.00	6.39
Cross-domain read $\rightarrow$ write (%)	0.00	0.98
Tool calls (matched)	100.0	100.0
Interaction length (relative)	100.0	99.1

Appendix

Table 6: Structural comparison between EES-Single and EES-Cross trajectories under length-controlled matching. Cross-domain metrics measure execution coordination beyond interaction length.

Domain	Selected operations	Declared requirements
Inventory	get_product_by_id ; update_product_stock	Non-negative prices; predefined product categories.
Promotions	get_product_by_sku ; get_applicable_promotions ; evaluate_customer_eligibility	Active time window, product–promotion link, and customer eligibility determine applicability.
Orders	get_product_by_id ; add_product_to_cart	Stock must cover the added quantity; each cart belongs to one user; repeated products use an aggregated quantity.

Appendix

Table 7: Selected interfaces in the recorded composition cross_env_0020 . Names are shown before namespace injection.

Step	Operation	Relevant observation or update
1	mobile_get_user_by_name	Resolves Maya Chen to usr-001 ; notification preference is push.
2	personal_get_note_by_id	Note content: compare service quotes and ask BioScope about January slots.
3	mobile_get_reminder_by_id	rem-001 belongs to usr-001 , with the existing message and time.
4	mobile_update_reminder_message	Sets “Call BioScope about January cryo-EM slots”; tool reports success.
5	mobile_update_reminder_time	Sets 2025-01-17 at 10:00, UTC $-$ 05:00; tool reports success.
6	mobile_get_user_notification_preference	Queries usr-001 ; result is push.

Appendix

Table 8: Model-visible tool evidence from one serialized two-domain EES-Cross demonstration. Construction-level states and verification records are retained separately from the conversation.

Source	Trajectories	Requests	Inventories	Declared tools	Invoked tools
EGTS	$13{,}824$	13,283	117	387	350
EES-Single	$8{,}026$	7,550	234	4,042	1,610
EES-Cross	$8{,}545$	8,319	276	4,316	1,683
Full SFT corpus	$30{,}395$	28,985	436	$4{,}672$	$2{,}260$

Appendix

Table 9: Coverage of the SFT corpus. “Requests” counts distinct normalized user-message sequences; “Inventories” counts distinct declared-tool configurations. Declared and invoked tools are counted by their serialized names.

Source	Mean	Min.	Max.	Mean	Min.	Max.	Mean	Min.	Max.
	Tokens			Tool calls			Assistant turns
EGTS	10,063.9	377	768,427	6.04	0	60	9.80	1	45
EES-Single	10,372.7	2,030	369,175	5.87	0	76	10.42	1	33
EES-Cross	20,444.7	1,746	1,259,991	7.01	0	99	7.57	1	34
Full SFT corpus	13,063.8	377	1,259,991	6.26	0	99	9.34	1	45

Appendix

Table 10: Length and interaction statistics of the SFT corpus. Token lengths describe complete serialized examples before training-time length handling.

Source	BFCL v4	$\boldsymbol{\tau^{2}}$ -bench	MCP-Atlas
EGTS	0.591	0.665	0.680
EES-Single	0.547	0.695	0.567
EES-Cross	0.575	0.670	0.589

Appendix

Table 11: Representation alignment between verified synthesis trajectories and benchmark reference or successful execution trajectories. Values are mean Top-5 cosine similarities under equal candidate budgets; bold marks the strongest source for each benchmark.

Benchmark	Matched	F $\rightarrow$ S	S $\rightarrow$ F	S $\rightarrow$ S	F $\rightarrow$ F
$\tau^{2}$ Retail	114	23	10	68	13
$\tau^{2}$ Telecom	113	74	4	35	0
BFCL v4	800	304	58	181	257
MCP-Atlas	89	40	6	9	34

Appendix

Table 12: Paired outcome transitions between the base model and Davinci-Tool-30B-SFT.

Panel of Figure 4 — Figure 4: Paired execution changes associated with benchmark recovery after SFT. (a) shows the fractions of matched tasks changing from failure to success and from success to failure. (b) shows that, among recovered $\tau^{2}$ -bench tasks, explicit tool-error rates decrease from 11.11% to 0.99% on Retail and from 12.72% to 0 on Telecom. (c) shows that recovered BFCL v4 trajectories become more complete: normalized reference-call coverage increases from 56.41% to 86.01%, visible output-to-argument dependency edges from 1.21 to 1.68, and visible chain depth from 1.58 to 1.79. (d) shows a different pattern on MCP-Atlas: observed empty retrievals decrease from 37 to 0 and protocol-fragment endings from 7 to 0, while calls per task increase from 5.55 to 8.90 and the explicit tool-error rate changes from 14.41% to 16.01%. The paired analysis therefore indicates benchmark-specific execution changes rather than a single mechanism based on shorter trajectories or uniformly fewer tool errors.

Setting	SFT (shared settings)	DPO
Optimizer	Fused AdamW	Fused AdamW
Learning rate	$1\times 10^{-5}$	$5\times 10^{-6}$
Epochs	4	1
Learning-rate schedule	Cosine	Cosine
Warmup fraction	0	0.03
Weight decay	0.01	0.01

Appendix

Table 13: Training hyperparameters. The SFT column applies to the 30B full-corpus model and the additional 8B, 14B, and Toucan-1.5M SFT runs. The DPO column applies to Davinci-Tool-30B. Batch size is per device, and gradient accumulation is measured in optimization steps.

LLM Tool Use

Large language model tool use studies how models select, call, and coordinate external tools to extend their reasoning and take actions.

See all 295 papers

Jun 11, 2026

HyperTool: Beyond Step-Wise Tool Calls for Tool-Augmented Agents

Tool-augmented LLM agents commonly rely on step-wise atomic tool calls, where each invocation, observation, and value transfer is exposed in the main reasoning trace. This creates an execution-granularity mismatch: locally deterministic tool workflows are unfolded into repeated model-visible decisions, consuming context and forcing the model to manage low-level dataflow in the trace. We introduce HyperTool, a unified executable MCP-style tool interface that changes the model-visible unit of tool execution. A model invokes HyperTool with a code block that can call existing tools through their original schemas, manipulate returned values, and pass intermediate results locally, folding deterministic tool subroutines into a single outer call. To train models to use this interface, we synthesize HyperTool-format trajectories from cross-tool compositional tasks and verify them in real MCP environments. On MCP-Universe, HyperTool improves average accuracy from 15.69% to 35.29% on Qwen3-32B and from 9.93% to 33.33% on Qwen3-8B, and surpass GPT-OSS and Kimi-k2.5 on average accuracy, showing that our HyperTool can substantially improve multi-step tool use.

Yaxin Du, Yifan Zhou, Yujie Ge +7

Shanghai Jiao Tong University · IQuest Research · Beijing University of Aeronautics and Astronautics

Tool-Augmented Language Model Agents LLM Agent Workflow Optimization LLM Tool Use LLM Agent Training Tool Use

Ask PDF

Sep 5, 2026

SAP: State-Guided Data Synthesis with Argument Provenance for Multi-Turn Tool Use

High-quality multi-turn tool-use data is essential for training agentic models, yet existing data synthesis methods often underrepresent the argument-level dependencies that are critical to long-horizon tool use. As a result, even when a model selects the correct tool, task execution may still fail because the model fills tool arguments with fabricated, stale, or weakly grounded values. To address this problem, we propose State-Guided Data Synthesis with Argument Provenance (SAP). SAP combines state guidance, tool-argument provenance constraints, and turn-level validation to efficiently construct tool-use trajectories with long-range dependencies and high accuracy. Using data generated by SAP, we build SAP-4B, which is highly competitive even when compared with much larger models across multiple benchmarks. Source code, synthesized data, and trained weights are available at https://github.com/Zichen1024/SAP.

Zichen Tian, Jinpeng Chen, Cheng Gong +2

Independent Researcher · Huawei Research

Synthetic Data Generation for Language Models LLM Agent Trajectory Synthesis LLM Agent Training Long-Horizon LLM Agents

Ask PDF

Mar 15, 2026

AgentProcessBench: Diagnosing Step-Level Process Quality in Tool-Using Agents

While Large Language Models (LLMs) have evolved into tool-using agents, they remain brittle in long-horizon interactions. Unlike mathematical reasoning where errors are often rectifiable via backtracking, tool-use failures frequently induce irreversible side effects, making accurate step-level verification critical. However, existing process-level benchmarks are predominantly confined to closed-world mathematical domains, failing to capture the dynamic and open-ended nature of tool execution. To bridge this gap, we introduce AgentProcessBench, the first benchmark dedicated to evaluating step-level effectiveness in realistic, tool-augmented trajectories. The benchmark comprises 1,000 diverse trajectories and 8,509 human-labeled step annotations with 89.1% inter-annotator agreement. It features a ternary labeling scheme to capture exploration and an error propagation rule to reduce labeling ambiguity. Extensive experiments reveal key insights: (1) weaker policy models exhibit inflated ratios of correct steps due to early termination; (2) distinguishing neutral and erroneous actions remains a significant challenge for current models; and (3) process-derived signals provide complementary value to outcome supervision, significantly enhancing test-time scaling. We hope AgentProcessBench can foster future research in reward models and pave the way toward general agents. The code and data are available at https://github.com/RUCBM/AgentProcessBench.

Shengda Fan, Xuyan Ye, Yupeng Huo +9

Renmin University of China Beijing, China · Beijing Jiaotong University Beijing, China · Shanghai Jiao Tong University Shanghai, China +1

LLM Agent Evaluation Process-Based Agent Evaluation Tool-Using Agents Agent Benchmarks Long-Horizon Agent Evaluation Process Supervision

Ask PDF

Abstract

LLM Tool Use

Explore similar work

HyperTool: Beyond Step-Wise Tool Calls for Tool-Augmented Agents

SAP: State-Guided Data Synthesis with Argument Provenance for Multi-Turn Tool Use

AgentProcessBench: Diagnosing Step-Level Process Quality in Tool-Using Agents