{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2LKAMCZKIQCTLJ5XNS4Z65SZYX","short_pith_number":"pith:2LKAMCZK","schema_version":"1.0","canonical_sha256":"d2d4060b2a440535a7b76cb99f7659c5c73f1b564d392e1dd47208d1f11727f5","source":{"kind":"arxiv","id":"2606.26607","version":1},"attestation_state":"computed","paper":{"title":"Moebius: Serving Mixture-of-Expert Models with Seamless Runtime Parallelism Switch","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Chong Li, Jaeyong Song, Seo Jin Park, Shaoyu Wang, Yizhuo Liang","submitted_at":"2026-06-25T05:10:20Z","abstract_excerpt":"Mixture-of-Experts (MoE) architectures scale large language models (LLMs) to hundreds of billions of parameters. Serving a single MoE model requires multiple GPUs operating in parallel, typically through tensor parallelism (TP) or expert parallelism (EP). The optimal choice depends on the number of in-flight requests: TP is faster at low concurrency, whereas EP wins at high concurrency. Production workloads cross this boundary continually: online serving sees bursty arrivals that subside into quiet periods, and reinforcement-learning rollouts begin as a high-concurrency burst that decays into "},"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":"2606.26607","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2026-06-25T05:10:20Z","cross_cats_sorted":[],"title_canon_sha256":"506bee9b8fd5f6c17ce27888615e0541cec796ae997f6e1fc046b0327b8a37d2","abstract_canon_sha256":"2cc82734d78138895003ef3d9aa2e980a9758a59ea160a81eecef01b0a682c3d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-26T01:15:36.088468Z","signature_b64":"Q0n0nt9WDRpGwYlvEo46jUMmz/y2+X//pb3DMeObOABcqXTICPXkwiXLLd97pSym2RFAN2+3H34Ik8GYCHsRCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d2d4060b2a440535a7b76cb99f7659c5c73f1b564d392e1dd47208d1f11727f5","last_reissued_at":"2026-06-26T01:15:36.088051Z","signature_status":"signed_v1","first_computed_at":"2026-06-26T01:15:36.088051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Moebius: Serving Mixture-of-Expert Models with Seamless Runtime Parallelism Switch","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Chong Li, Jaeyong Song, Seo Jin Park, Shaoyu Wang, Yizhuo Liang","submitted_at":"2026-06-25T05:10:20Z","abstract_excerpt":"Mixture-of-Experts (MoE) architectures scale large language models (LLMs) to hundreds of billions of parameters. Serving a single MoE model requires multiple GPUs operating in parallel, typically through tensor parallelism (TP) or expert parallelism (EP). The optimal choice depends on the number of in-flight requests: TP is faster at low concurrency, whereas EP wins at high concurrency. Production workloads cross this boundary continually: online serving sees bursty arrivals that subside into quiet periods, and reinforcement-learning rollouts begin as a high-concurrency burst that decays into "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26607","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/2606.26607/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":"2606.26607","created_at":"2026-06-26T01:15:36.088105+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.26607v1","created_at":"2026-06-26T01:15:36.088105+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26607","created_at":"2026-06-26T01:15:36.088105+00:00"},{"alias_kind":"pith_short_12","alias_value":"2LKAMCZKIQCT","created_at":"2026-06-26T01:15:36.088105+00:00"},{"alias_kind":"pith_short_16","alias_value":"2LKAMCZKIQCTLJ5X","created_at":"2026-06-26T01:15:36.088105+00:00"},{"alias_kind":"pith_short_8","alias_value":"2LKAMCZK","created_at":"2026-06-26T01:15:36.088105+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/2LKAMCZKIQCTLJ5XNS4Z65SZYX","json":"https://pith.science/pith/2LKAMCZKIQCTLJ5XNS4Z65SZYX.json","graph_json":"https://pith.science/api/pith-number/2LKAMCZKIQCTLJ5XNS4Z65SZYX/graph.json","events_json":"https://pith.science/api/pith-number/2LKAMCZKIQCTLJ5XNS4Z65SZYX/events.json","paper":"https://pith.science/paper/2LKAMCZK"},"agent_actions":{"view_html":"https://pith.science/pith/2LKAMCZKIQCTLJ5XNS4Z65SZYX","download_json":"https://pith.science/pith/2LKAMCZKIQCTLJ5XNS4Z65SZYX.json","view_paper":"https://pith.science/paper/2LKAMCZK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.26607&json=true","fetch_graph":"https://pith.science/api/pith-number/2LKAMCZKIQCTLJ5XNS4Z65SZYX/graph.json","fetch_events":"https://pith.science/api/pith-number/2LKAMCZKIQCTLJ5XNS4Z65SZYX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2LKAMCZKIQCTLJ5XNS4Z65SZYX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2LKAMCZKIQCTLJ5XNS4Z65SZYX/action/storage_attestation","attest_author":"https://pith.science/pith/2LKAMCZKIQCTLJ5XNS4Z65SZYX/action/author_attestation","sign_citation":"https://pith.science/pith/2LKAMCZKIQCTLJ5XNS4Z65SZYX/action/citation_signature","submit_replication":"https://pith.science/pith/2LKAMCZKIQCTLJ5XNS4Z65SZYX/action/replication_record"}},"created_at":"2026-06-26T01:15:36.088105+00:00","updated_at":"2026-06-26T01:15:36.088105+00:00"}