{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/tora-a-tool-integrated-reasoning-agent-for","title":"ToRA: A Tool-Integrated Reasoning Agent for Mathematical Problem Solving","arxiv_id":"2309.17452","date":"2023-09-29","proceeding":null,"authors":["Zhibin Gou","Zhihong Shao","Yeyun Gong","Yelong Shen","Yujiu Yang","Minlie Huang","Nan Duan","Weizhu Chen"],"abstract":"Large language models have made significant progress in various language tasks, yet they still struggle with complex mathematics. In this paper, we propose ToRA a series of Tool-integrated Reasoning Agents designed to solve challenging mathematical problems by seamlessly integrating natural language reasoning with the utilization of external tools (e.g., computation libraries and symbolic solvers), thereby amalgamating the analytical prowess of language and the computational efficiency of tools. To train ToRA, we curate interactive tool-use trajectories on mathematical datasets, apply imitation learning on the annotations, and propose output space shaping to further refine models' reasoning behavior. As a result, ToRA models significantly outperform open-source models on 10 mathematical reasoning datasets across all scales with 13%-19% absolute improvements on average. Notably, ToRA-7B reaches 44.6% on the competition-level dataset MATH, surpassing the best open-source model WizardMath-70B by 22% absolute. ToRA-Code-34B is also the first open-source model that achieves an accuracy exceeding 50% on MATH, which significantly outperforms GPT-4's CoT result, and is competitive with GPT-4 solving problems with programs. Additionally, we conduct a comprehensive analysis of the benefits and remaining challenges of tool interaction for mathematical reasoning, providing valuable insights for future research.","url_abs":"https://arxiv.org/abs/2309.17452v4","url_pdf":"https://arxiv.org/pdf/2309.17452v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"tora-a-tool-integrated-reasoning-agent-for","repo_url":"https://github.com/microsoft/tora","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"arithmetic-reasoning","task_name":"Arithmetic Reasoning"},{"task_slug":"computational-efficiency","task_name":"Computational Efficiency"},{"task_slug":"imitation-learning","task_name":"Imitation Learning"},{"task_slug":"math","task_name":"Math"},{"task_slug":"math-word-problem-solving","task_name":"Math Word Problem Solving"},{"task_slug":"mathematical-problem-solving","task_name":"Mathematical Problem-Solving"},{"task_slug":"mathematical-reasoning","task_name":"Mathematical Reasoning"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"gpt-4","method_name":"GPT-4"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/arithmetic-reasoning-on-gsm8k","task":"Arithmetic Reasoning","dataset":"GSM8K","model":"ToRA-70B (SC, k=50)","rank_in_archive_order":28,"of":164,"metrics":{"Accuracy":"88.3","Parameters (Billion)":"70"},"uses_additional_data":true},{"leaderboard":"/sota/arithmetic-reasoning-on-gsm8k","task":"Arithmetic Reasoning","dataset":"GSM8K","model":"ToRA-Code-34B (SC, k=50)","rank_in_archive_order":45,"of":164,"metrics":{"Accuracy":"85.1","Parameters (Billion)":"34"},"uses_additional_data":true},{"leaderboard":"/sota/arithmetic-reasoning-on-gsm8k","task":"Arithmetic Reasoning","dataset":"GSM8K","model":"ToRA 70B","rank_in_archive_order":51,"of":164,"metrics":{"Accuracy":"84.3","Parameters (Billion)":"70"},"uses_additional_data":true},{"leaderboard":"/sota/arithmetic-reasoning-on-gsm8k","task":"Arithmetic Reasoning","dataset":"GSM8K","model":"ToRA-Code 34B","rank_in_archive_order":69,"of":164,"metrics":{"Accuracy":"80.7","Parameters (Billion)":"34"},"uses_additional_data":true},{"leaderboard":"/sota/arithmetic-reasoning-on-gsm8k","task":"Arithmetic Reasoning","dataset":"GSM8K","model":"ToRA-Code 13B","rank_in_archive_order":85,"of":164,"metrics":{"Accuracy":"75.8","Parameters (Billion)":"13"},"uses_additional_data":true},{"leaderboard":"/sota/arithmetic-reasoning-on-gsm8k","task":"Arithmetic Reasoning","dataset":"GSM8K","model":"ToRA-Code 7B","rank_in_archive_order":98,"of":164,"metrics":{"Accuracy":"72.6","Parameters (Billion)":"7"},"uses_additional_data":true},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"ToRA-Code 34B model (w/ code, SC, k=50)","rank_in_archive_order":27,"of":135,"metrics":{"Accuracy":"60.0","Parameters (Billions)":"34"},"uses_additional_data":true},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"ToRA 70B (w/ code, SC, k=50)","rank_in_archive_order":33,"of":135,"metrics":{"Accuracy":"56.9","Parameters (Billions)":"70"},"uses_additional_data":true},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"ToRA-Code 34B (w/ code)","rank_in_archive_order":46,"of":135,"metrics":{"Accuracy":"50.8","Parameters (Billions)":"34"},"uses_additional_data":true},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"ToRA 70B (w/ code)","rank_in_archive_order":49,"of":135,"metrics":{"Accuracy":"49.7","Parameters (Billions)":"70"},"uses_additional_data":true},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"ToRA-Code 13B (w/ code)","rank_in_archive_order":56,"of":135,"metrics":{"Accuracy":"48.1","Parameters (Billions)":"13"},"uses_additional_data":true},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"ToRA-Code 7B (w/ code)","rank_in_archive_order":67,"of":135,"metrics":{"Accuracy":"44.6","Parameters (Billions)":"7"},"uses_additional_data":true},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"ToRA 13B (w/ code)","rank_in_archive_order":74,"of":135,"metrics":{"Accuracy":"43.0","Parameters (Billions)":"13"},"uses_additional_data":true},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"ToRA 7B (w/ code)","rank_in_archive_order":78,"of":135,"metrics":{"Accuracy":"40.1","Parameters (Billions)":"7"},"uses_additional_data":true}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2309.17452","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2309.17452"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/microsoft/tora","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_draft_wrong":2,"ran":3,"unverified":7},"by_repo_kind":{"official":{"samples":12,"ran":5,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"bc1a96ec28a11a5f","entry":"encode_with_messages_format","repo":"microsoft/ToRA","repo_kind":"official","path":"src/train/finetune.py","file_url":"https://github.com/microsoft/ToRA/blob/HEAD/src/train/finetune.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"bc1a96ec28a11a5f"}},{"code_sha256_prefix":"1906977026be9a86","entry":"encode_with_prompt_completion_format","repo":"microsoft/ToRA","repo_kind":"official","path":"src/train/finetune.py","file_url":"https://github.com/microsoft/ToRA/blob/HEAD/src/train/finetune.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1906977026be9a86"}},{"code_sha256_prefix":"8ae8c583b21633b3","entry":"is_digit","repo":"microsoft/tora","repo_kind":"official","path":"src/eval/grader.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/eval/grader.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8ae8c583b21633b3"}},{"code_sha256_prefix":"65c738adb08e913c","entry":"math_equal","repo":"microsoft/tora","repo_kind":"official","path":"src/eval/grader.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/eval/grader.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"65c738adb08e913c"}},{"code_sha256_prefix":"f63c83de2a26f67c","entry":"math_equal_process","repo":"microsoft/tora","repo_kind":"official","path":"src/eval/grader.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/eval/grader.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f63c83de2a26f67c"}},{"code_sha256_prefix":"3408dd47e939d760","entry":"apply_rotary_pos_emb","repo":"microsoft/tora","repo_kind":"official","path":"src/train/llama2_flash_attn_monkey_patch.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/train/llama2_flash_attn_monkey_patch.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3408dd47e939d760"}},{"code_sha256_prefix":"ecdacf069c9d10db","entry":"construct_prompt","repo":"microsoft/tora","repo_kind":"official","path":"src/utils/utils.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/utils/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ecdacf069c9d10db"}},{"code_sha256_prefix":"951c55f8667c735f","entry":"extract_answer","repo":"microsoft/tora","repo_kind":"official","path":"src/utils/parser.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/utils/parser.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"951c55f8667c735f"}},{"code_sha256_prefix":"1d961161c71cc153","entry":"extract_program","repo":"microsoft/tora","repo_kind":"official","path":"src/utils/parser.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/utils/parser.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1d961161c71cc153"}},{"code_sha256_prefix":"3bb02b5083de68fd","entry":"load_prompt","repo":"microsoft/tora","repo_kind":"official","path":"src/utils/utils.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/utils/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3bb02b5083de68fd"}},{"code_sha256_prefix":"0af0b06d6cf22524","entry":"lower_keys","repo":"microsoft/tora","repo_kind":"official","path":"src/utils/utils.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/utils/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0af0b06d6cf22524"}},{"code_sha256_prefix":"06b04cb075356fb9","entry":"strip_string","repo":"microsoft/tora","repo_kind":"official","path":"src/utils/parser.py","file_url":"https://github.com/microsoft/tora/blob/HEAD/src/utils/parser.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"06b04cb075356fb9"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}