{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:URXOONOMMNQQ56BP5RVBGS5UYI","short_pith_number":"pith:URXOONOM","schema_version":"1.0","canonical_sha256":"a46ee735cc63610ef82fec6a134bb4c205ea466e4eac585db630f7c10c43fcf8","source":{"kind":"arxiv","id":"2608.16157","version":1},"attestation_state":"computed","paper":{"title":"FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Chenfeng Xu, Haocheng Xi, Ion Stoica, Kurt Keutzer, Matei Zaharia, Melissa Pan, Shanlin Sun, Shuo Yang, Song Han, Xiaoze Fan, Zhe Wang","submitted_at":"2026-08-17T06:22:53Z","abstract_excerpt":"Frontier open-weight models are increasingly available, but serving them still largely assumes datacenter infrastructure. We present FreeToken, an edge-native MoE serving system that treats a personal machine not as a small GPU, but as a unified, elastic inference platform. FreeToken co-designs the full serving stack, including model layout and loading, expert residency, CPU--GPU execution, agentic state reuse, and runtime memory management, around two realities of local AI: agent workloads continuously change their execution pattern, and edge hardware exposes heterogeneous resources whose bal"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2608.16157","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2026-08-17T06:22:53Z","cross_cats_sorted":[],"title_canon_sha256":"194cbb04c241d065019863868720aa2bf80193f438c804f2ac49e699b18e68b4","abstract_canon_sha256":"1d711ea607d1156ea298d4f0dc58f57049fedd2dabf45da3855b889d6a8055d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-18T02:17:22.927550Z","signature_b64":"iyZNYJ3O0cIVxVVDQi21B5v+AZnwgwm6Va/kD2Y6DclI2ArB7iCqkbxVPtlimQTX1DwjqIr+jTncChHMVaUqAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a46ee735cc63610ef82fec6a134bb4c205ea466e4eac585db630f7c10c43fcf8","last_reissued_at":"2026-08-18T02:17:22.926041Z","signature_status":"signed_v1","first_computed_at":"2026-08-18T02:17:22.926041Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Chenfeng Xu, Haocheng Xi, Ion Stoica, Kurt Keutzer, Matei Zaharia, Melissa Pan, Shanlin Sun, Shuo Yang, Song Han, Xiaoze Fan, Zhe Wang","submitted_at":"2026-08-17T06:22:53Z","abstract_excerpt":"Frontier open-weight models are increasingly available, but serving them still largely assumes datacenter infrastructure. We present FreeToken, an edge-native MoE serving system that treats a personal machine not as a small GPU, but as a unified, elastic inference platform. FreeToken co-designs the full serving stack, including model layout and loading, expert residency, CPU--GPU execution, agentic state reuse, and runtime memory management, around two realities of local AI: agent workloads continuously change their execution pattern, and edge hardware exposes heterogeneous resources whose bal"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.16157","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2608.16157/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2608.16157","created_at":"2026-08-18T02:17:22.925627+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.16157v1","created_at":"2026-08-18T02:17:22.925627+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.16157","created_at":"2026-08-18T02:17:22.925627+00:00"},{"alias_kind":"pith_short_12","alias_value":"URXOONOMMNQQ","created_at":"2026-08-18T02:17:22.925627+00:00"},{"alias_kind":"pith_short_16","alias_value":"URXOONOMMNQQ56BP","created_at":"2026-08-18T02:17:22.925627+00:00"},{"alias_kind":"pith_short_8","alias_value":"URXOONOM","created_at":"2026-08-18T02:17:22.925627+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI","json":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI.json","graph_json":"https://pith.science/api/pith-number/URXOONOMMNQQ56BP5RVBGS5UYI/graph.json","events_json":"https://pith.science/api/pith-number/URXOONOMMNQQ56BP5RVBGS5UYI/events.json","paper":"https://pith.science/paper/URXOONOM"},"agent_actions":{"view_html":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI","download_json":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI.json","view_paper":"https://pith.science/paper/URXOONOM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.16157&json=true","fetch_graph":"https://pith.science/api/pith-number/URXOONOMMNQQ56BP5RVBGS5UYI/graph.json","fetch_events":"https://pith.science/api/pith-number/URXOONOMMNQQ56BP5RVBGS5UYI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI/action/storage_attestation","attest_author":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI/action/author_attestation","sign_citation":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI/action/citation_signature","submit_replication":"https://pith.science/pith/URXOONOMMNQQ56BP5RVBGS5UYI/action/replication_record"}},"created_at":"2026-08-18T02:17:22.925627+00:00","updated_at":"2026-08-18T02:17:22.925627+00:00"}