{ "schema_version": 1, "builder_plan": { "scenario": "An assistant chooses the wrong API or sends arguments your product cannot use.", "change": "Apply the proposed tool-selection or argument-generation change in a sandbox.", "baseline": "Your current agent with the same tool schemas and task budget.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Correct completed actions, invalid calls, retries, latency, and cost per completed task.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2609.21967", "title": "NemotronLabs VoiceChat: An Open Full-duplex Speech-to-Speech Model with Tool Calling Capabilities", "summary": "The technique replaces the conventional cascaded voice-agent pipeline (streaming ASR, LLM with tool calling, and TTS chained as separate services) with a single unified full-duplex speech-to-speech model that handles interruption, backchanneling, and function calls natively. It costs the deployment burden of a multi-component streaming model (speech encoder, decoder LM, RNN-T branch, streaming TTS decoder) whose total footprint is not stated in the abstract and may exceed a single 24 GB GPU, plus the engineering cost of integrating streaming audio I/O and tool runtimes. It can fail on tool execution specifically, since the authors state argument accuracy and end-to-end tool execution remain areas for improvement, and any regression in full-duplex timing degrades conversational quality in ways the text-only stack would not.", "rationale": "The reader's constrained infrastructure and production focus make a newly released open full-duplex speech model with unverified local resource requirements and known tool-execution weaknesses something to track rather than deploy or validate immediately.", "technique": "Full-duplex speech-to-speech model with native tool calling", "family": "uso_de_ferramenta", "recommendation": "observar", "published": "2026-09-18", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.21967", "independent_implementations": 3, "total_implementations": 4, "stars": 125, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "The technique replaces the conventional cascaded voice-agent pipeline (streaming ASR, LLM with tool calling, and TTS chained as separate services) with a single unified full-duplex speech-to-speech model that handles interruption, backchanneling, and function calls natively. It costs the deployment burden of a multi-component streaming model (speech encoder, decoder LM, RNN-T branch, streaming TTS decoder) whose total footprint is not stated in the abstract and may exceed a single 24 GB GPU, plus the engineering cost of integrating streaming audio I/O and tool runtimes. It can fail on tool execution specifically, since the authors state argument accuracy and end-to-end tool execution remain areas for improvement, and any regression in full-duplex timing degrades conversational quality in ways the text-only stack would not.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.21967", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "On Full-Duplex-Bench 3.0, 82.5% tool-selection F1; 55.1 normalized average on VoiceBench; lowest pause-handling takeover rates among open-weight systems, with argument accuracy and end-to-end tool execution noted as weak points.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.21967", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }