{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:LSHC4VJQIPVQGGUU4WDFNWYPXK","short_pith_number":"pith:LSHC4VJQ","canonical_record":{"source":{"id":"2607.20468","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-20T15:55:21Z","cross_cats_sorted":[],"title_canon_sha256":"c6572524f9a4e7382e16c47cd6e844922e7a720455b48532ebc715aac0f0270c","abstract_canon_sha256":"7e2a1d370366f3ce2c8acda36e26a4a9052fefb752e0da95358a82ec8bb87802"},"schema_version":"1.0"},"canonical_sha256":"5c8e2e553043eb031a94e58656db0fbab4082af9b4b7b29c427bd47be5d9917c","source":{"kind":"arxiv","id":"2607.20468","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.20468","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"arxiv_version","alias_value":"2607.20468v1","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.20468","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"pith_short_12","alias_value":"LSHC4VJQIPVQ","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"pith_short_16","alias_value":"LSHC4VJQIPVQGGUU","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"pith_short_8","alias_value":"LSHC4VJQ","created_at":"2026-07-24T00:23:17Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:LSHC4VJQIPVQGGUU4WDFNWYPXK","target":"record","payload":{"canonical_record":{"source":{"id":"2607.20468","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-20T15:55:21Z","cross_cats_sorted":[],"title_canon_sha256":"c6572524f9a4e7382e16c47cd6e844922e7a720455b48532ebc715aac0f0270c","abstract_canon_sha256":"7e2a1d370366f3ce2c8acda36e26a4a9052fefb752e0da95358a82ec8bb87802"},"schema_version":"1.0"},"canonical_sha256":"5c8e2e553043eb031a94e58656db0fbab4082af9b4b7b29c427bd47be5d9917c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-24T00:23:17.296571Z","signature_b64":"V0ONOyTXWjm/BcqGEvUF0tEX6CAXRrUZzX6N9qscFEGHpEhBtcBANi2qlOMM+lo+FQp3m+eADJZbkzkWSlP7DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5c8e2e553043eb031a94e58656db0fbab4082af9b4b7b29c427bd47be5d9917c","last_reissued_at":"2026-07-24T00:23:17.295685Z","signature_status":"signed_v1","first_computed_at":"2026-07-24T00:23:17.295685Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2607.20468","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-24T00:23:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"AibSV3CrecL0xZ6tcGyQkDGxjIulNW7ov9GE3VhwuYHjMVVlfquvOK9JBghUe8WlMZEMNnU4bFuKGPTfsI58Cw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T11:57:45.093591Z"},"content_sha256":"6ec9620b134a9da5ab902975ecda28f499c7d93286d8a8169a5c3b68c0cf309e","schema_version":"1.0","event_id":"sha256:6ec9620b134a9da5ab902975ecda28f499c7d93286d8a8169a5c3b68c0cf309e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:LSHC4VJQIPVQGGUU4WDFNWYPXK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"InferenceBench: A Benchmark for Open-Ended LLM Inference Optimization by AI Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Ben Rank, Jehyeok Yeon, Maksym Andriushchenko","submitted_at":"2026-05-20T15:55:21Z","abstract_excerpt":"AI agents are increasingly used to automate research and development tasks, yet existing benchmarks typically evaluate them on prescribed workflows or narrow action spaces. Even nominally open-ended tasks can often be solved by retrieving a well-known recipe and tuning a few hyperparameters, making it unclear whether strong results reflect genuine optimization or memorized solutions. We introduce InferenceBench, where an agent must deploy an OpenAI-compatible inference server and optimize the speed of LLM inference. Each agent receives a target LLM, one H100 GPU, an optimization scenario, and "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.20468","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/2607.20468/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-24T00:23:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PrB4TiTSiyG1Ul3AUmBd5sZ0PwHgf2iWATSZ+XfX68eC/1NTB5pYXrIF8d7kFZEcgtS5QxymnYtQf9ox+ChwAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T11:57:45.094086Z"},"content_sha256":"3ef3d4f004388121be176979445919667bfa7a1e61e6f52c8af67fde8a96b161","schema_version":"1.0","event_id":"sha256:3ef3d4f004388121be176979445919667bfa7a1e61e6f52c8af67fde8a96b161"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LSHC4VJQIPVQGGUU4WDFNWYPXK/bundle.json","state_url":"https://pith.science/pith/LSHC4VJQIPVQGGUU4WDFNWYPXK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LSHC4VJQIPVQGGUU4WDFNWYPXK/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-04T11:57:45Z","links":{"resolver":"https://pith.science/pith/LSHC4VJQIPVQGGUU4WDFNWYPXK","bundle":"https://pith.science/pith/LSHC4VJQIPVQGGUU4WDFNWYPXK/bundle.json","state":"https://pith.science/pith/LSHC4VJQIPVQGGUU4WDFNWYPXK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LSHC4VJQIPVQGGUU4WDFNWYPXK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LSHC4VJQIPVQGGUU4WDFNWYPXK","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"7e2a1d370366f3ce2c8acda36e26a4a9052fefb752e0da95358a82ec8bb87802","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-20T15:55:21Z","title_canon_sha256":"c6572524f9a4e7382e16c47cd6e844922e7a720455b48532ebc715aac0f0270c"},"schema_version":"1.0","source":{"id":"2607.20468","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.20468","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"arxiv_version","alias_value":"2607.20468v1","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.20468","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"pith_short_12","alias_value":"LSHC4VJQIPVQ","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"pith_short_16","alias_value":"LSHC4VJQIPVQGGUU","created_at":"2026-07-24T00:23:17Z"},{"alias_kind":"pith_short_8","alias_value":"LSHC4VJQ","created_at":"2026-07-24T00:23:17Z"}],"graph_snapshots":[{"event_id":"sha256:3ef3d4f004388121be176979445919667bfa7a1e61e6f52c8af67fde8a96b161","target":"graph","created_at":"2026-07-24T00:23:17Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2607.20468/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"AI agents are increasingly used to automate research and development tasks, yet existing benchmarks typically evaluate them on prescribed workflows or narrow action spaces. Even nominally open-ended tasks can often be solved by retrieving a well-known recipe and tuning a few hyperparameters, making it unclear whether strong results reflect genuine optimization or memorized solutions. We introduce InferenceBench, where an agent must deploy an OpenAI-compatible inference server and optimize the speed of LLM inference. Each agent receives a target LLM, one H100 GPU, an optimization scenario, and ","authors_text":"Ben Rank, Jehyeok Yeon, Maksym Andriushchenko","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-20T15:55:21Z","title":"InferenceBench: A Benchmark for Open-Ended LLM Inference Optimization by AI Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.20468","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6ec9620b134a9da5ab902975ecda28f499c7d93286d8a8169a5c3b68c0cf309e","target":"record","created_at":"2026-07-24T00:23:17Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"7e2a1d370366f3ce2c8acda36e26a4a9052fefb752e0da95358a82ec8bb87802","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-20T15:55:21Z","title_canon_sha256":"c6572524f9a4e7382e16c47cd6e844922e7a720455b48532ebc715aac0f0270c"},"schema_version":"1.0","source":{"id":"2607.20468","kind":"arxiv","version":1}},"canonical_sha256":"5c8e2e553043eb031a94e58656db0fbab4082af9b4b7b29c427bd47be5d9917c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5c8e2e553043eb031a94e58656db0fbab4082af9b4b7b29c427bd47be5d9917c","first_computed_at":"2026-07-24T00:23:17.295685Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-24T00:23:17.295685Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"V0ONOyTXWjm/BcqGEvUF0tEX6CAXRrUZzX6N9qscFEGHpEhBtcBANi2qlOMM+lo+FQp3m+eADJZbkzkWSlP7DQ==","signature_status":"signed_v1","signed_at":"2026-07-24T00:23:17.296571Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.20468","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6ec9620b134a9da5ab902975ecda28f499c7d93286d8a8169a5c3b68c0cf309e","sha256:3ef3d4f004388121be176979445919667bfa7a1e61e6f52c8af67fde8a96b161"],"state_sha256":"7706fe2159996064252f2703900b74d76784ceaa10bfe90cedbbd96e6df5b185"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kdMmhKTtXPk7EZGWVOeZmamXANmNssroCYJhx4vfieOlD6HrmLZ6ODv12xae9jIcp8ClzBS447huMDsgqF51Bg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T11:57:45.098990Z","bundle_sha256":"0edc9ccdf53f0a2692dee340e8d4157140f4924909d6fef11971ec255a6c0889"}}