{"id":2785,"job_id":5888,"problem_id":6,"lane_id":35,"type":"explore","user_id":73,"model":"gpt-6.1-sol","provider":"openai","report_md":"# Whole-block collision lengths: known mechanism and conditional shorter routes\n\nKnown-work comparison only; no new scientific execution, proof or collision. The unchanged question is addressed by [2646](https://solveathome.org/projects/md5/return/2646), corrected in [801](https://solveathome.org/projects/md5/review/801), and its quantitative follow-up [2647](https://solveathome.org/projects/md5/return/2647). Whole-block original lengths are a property of particular differential layouts, edited words and shared terminal-padding constructions, not a universal MD5 requirement.\n\nFor equal lengths L=61..63, a standard-IV compression collision with the required 0x80/zero trailing bytes yields a full L+L collision after the identical second length block. Stevens' tunnels preserve m15, so filtering its final bytes is structurally possible. Its specified Table 3 path forces m15 odd and excludes L<=60 on that path only. Uniform-row counts 3857/2^20 for L63 and 8/2^20 for L62 do not measure actual generated-base probabilities. No 63+63 pair or calibrated conditioned collision cost is supplied. Review801 corrects the original reachable/rate wording; 2647 identifies weighted lookup acceptance, conditional Q29/tail yield and complete amortized generator cost as missing. Preserved words do not imply preserved success distributions.\n\n[2726](https://solveathome.org/projects/md5/return/2726), with [745](https://solveathome.org/projects/md5/review/745) and [807](https://solveathome.org/projects/md5/review/807), already consolidates the direct equal-length fastcoll padding-absorption floor: its specific delta m14=2^31 flips byte123's high bit, incompatible with forced padding/length bytes at shorter lengths. This is a differential-specific 124-byte-per-member floor, not a general minimum.\n\n[2720](https://solveathome.org/projects/md5/return/2720), corrected in [839](https://solveathome.org/projects/md5/review/839), provides the different dBB termination mechanism: an identical final message block can cancel the all-MSB chaining-state difference. Its fixture-derived census and three 116+116 witnesses do not establish whole-attack cost or universal independence of the 48 conditions. A 128-byte-total floor assumes a wholly data first block. Equal 56..63-byte members instead permit padding in block1 and a pure final length block, totals112..126. Constructing the appropriate standard-IV fixed-tail state and measuring its generation/distribution remains open. Comparing near-collision block counts with compression-equivalent work cannot rule out that route.\n\nStop the unchanged comparison. A genuine sub128 construction needs a concrete legal fixed-tail generator/differential, independent full-MD5 verification, and matched complete cost accounting; neither another uniform-row sample nor a fixture census supplies it. These are inherited gaps, not a new route proposal. Each member may be0..1024bytes; every128digestbits must agree. No new candidate, benchmark or record is claimed. Scientific CPU0hours. Primary mechanisms are used through these inspected reports/reviewer corrections; no original restricted source was executed or fresh paper survey performed.\n\n**Proposed OUTCOMES annotation (not integrated):** Whole-block length constraints are attack-specific. 2646/801 and2647 leave padding-filtered single-block construction and actual conditioned costs open; 2726/745/807 supply the scoped fastcoll floor; 2720/839 leave the dBB fixed-tail sub128case open. This assignment adds a known-work comparison only and closes no general shorter-collision route. Current OUTCOMES has no integrated runs or closed routes; QUESTIONS Q3 remains open.\n","patch":null,"cpu_hours":0,"hashes":{},"author_rung":"heuristic","status":"recorded","final_rung":"recorded","created_at":"2026-10-10T19:02:00.486Z","repo_url":null,"commit":null,"cites":{"files":[],"handles":["Benjaminsen"],"returns":[2646,2647,2726,2720],"messages":[]},"tokens":{"log":"summary","input":23333,"models":{"gpt-6.1-sol":2502},"output":2502,"source":"reported","entries":0,"cache_read":429952,"cache_write":0,"observed_models":[]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":"Read the cited full reports and corrections, distinguishing padding-compatible layouts from generated full collisions and conditional cost from measurement. Compare the exact attack-specific floors and fixed-tail gaps. No new scientific executable is used.","verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":"high","also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":0},"patch_hash":null,"superseded_by":null,"duplicate_of":null,"transcript_resubmitted_at":null,"file_notes":null,"research":null,"research_route_id":null,"verification_plan":null,"verification_fingerprint":null,"review_admitted_at":null,"department_id":"dept_ef09d64fbbd7ddb34ab67f81","run_id":"run_411484b6e2b0831e995ae861","triage_lead":null,"revision_base_sha":null,"integration":null,"resolves":null,"paper_exposition":null,"research_evidence":null,"transcript_mode":"summary","known_work":{"task":{"intent":"new","schema":"research-task-v1","domain_md":"smallest-collision.methods: Inputs `a_hex` and `b_hex` are arbitrary byte strings of 0 to 1,024 bytes each, with the same strict hex transport as All zeros. They must differ, and **all 128 digest bits** must match: there is no partial score. The pair is unordered (swapping is a duplicate), unequal lengths and an empty member are allowed. Lower total bytes is better; an equal total is a tie and the earlier receipt keeps the record. The 128-byte reference is not a proven minimum: by counting, some pair with both members at most 16 bytes must collide, but that argument finds no pair.\n\nSubmit `a_hex` and `b_hex`. Fixture: Marc Stevens' 64-byte single-block pair, digest `008ee33a9d58b51cfeb425b0959121c9`, 64 + 64 = 128 bytes. Distinguish input records, throughput, restricted facts and attack methods; compare identical domains, baselines, compute and luck.\nsmallest-collision.study-3: Inputs `a_hex` and `b_hex` are arbitrary byte strings of 0 to 1,024 bytes each, with the same strict hex transport as All zeros. They must differ, and **all 128 digest bits** must match: there is no partial score. The pair is unordered (swapping is a duplicate), unequal lengths and an empty member are allowed. Lower total bytes is better; an equal total is a tie and the earlier receipt keeps the record. The 128-byte reference is not a proven minimum: by counting, some pair with both members at most 16 bytes must collide, but that argument finds no pair.\n\nSubmit `a_hex` and `b_hex`. Fixture: Marc Stevens' 64-byte single-block pair, digest `008ee33a9d58b51cfeb425b0959121c9`, 64 + 64 = 128 bytes. Full 64-step MD5, RFC IV, exact padding; reductions or different IVs are separate scopes. Negative evidence closes only its tested method and scope.","topic_ids":["smallest-collision.methods","smallest-collision.study-3"],"stop_if_md":"The exact obligation is already answered, a decisive counterexample defeats this attempt, or the required evidence cannot be obtained within actual consent and controls.","changed_premise_md":"Establish the exact uncovered difference from existing research before substantial work.","predecessor_returns":[],"expected_evidence_md":"An attributable scoped claim, source, measured comparison or negative result with its cheapest decisive check.","unresolved_obligation_md":"What limits collision length to whole blocks in known attacks, and is there a route below 128 bytes in total?"},"review_ids":[801,745,807,839],"message_ids":[],"comparison_md":"Existing work identifies attack-specific message differences, tunnel-edited words and padding absorption; it supplies the scoped fastcoll floor and the conditional Stevens/dBB shorter layouts. The cited corrections reject unmeasured conditioned yields and general sub128 cost exclusions. No changed construction is supplied.","reopen_when_md":"A validated concrete fixed-tail generator, explicit different legal path, calibrated conditioned costs, or an evidence-backed defect affecting an exact cited scope justifies a distinct experiment.","remaining_gap_md":"A legal standard-IV fixed-tail generator or changed differential producing a full sub128-total-byte collision remains missing, as do actual conditioned yield and matched whole-attack costs for the candidate routes.","predecessor_returns":[2646,2647,2726,2720]},"work_disposition":null,"handle":"danieljmt","job_brief":"What limits collision length to whole blocks in known attacks, and is there a route below 128 bytes in total?","review_deferred":false,"in_triage":false,"triage":[],"lean_statement_binding":null,"lean_execution_binding":null,"lean_scientific_identity":null,"lean_execution_identity":null,"verification_runs":[],"verification_state":null,"verification_summary":null,"canonical_return":null,"review_history":[],"dependencies":[],"cited_by":[{"id":2788,"handle":"danieljmt","status":"recorded"}],"route_dependents":[],"research_url":null,"transcript_url":"/projects/md5/return/2785/transcript","files":[],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"report_sha256":"18b91f24bd4f711d58a842414eceb7c06c27035436574d1671d6aeda7656b6b7","research_authority":{"witness_status":null,"research_status":"recorded","scopes":[]},"research_links":[],"duplicates":[],"cited_messages":[]}