tool,trial,test_date,account_tier_mode,completion_status,reported_result_count,accessible_result_count,full_text_count,claim_source_checks,failure_flag,access_barrier,elapsed_minutes,evidence_url,observation
ResearchRabbit,1,2026-08-01,Free / keyword search,Complete,18800,,,0,false,,0.2,https://app.researchrabbit.ai/search/a344a73f-013c-4aac-b771-e0dbbb6c863e/0,Broad discovery result; reviews and non-comparators were mixed with plausible primary studies.
ResearchRabbit,2,2026-08-01,Free / keyword search,Complete,2390,,,0,false,,0.2,https://app.researchrabbit.ai/search/acb6d1e9-889e-454a-8c1c-e12ccb3131c2/0,Identical keywords returned far fewer results and materially different top rankings.
ResearchRabbit,3,2026-08-01,Free / keyword search,Complete,18900,,,0,false,,0.2,https://app.researchrabbit.ai/search/ade6b526-be21-46e4-8047-586b70e40775/0,"Result volume reverted close to trial 1, confirming a large between-run count and ranking swing."
Elicit,1,2026-08-01,Free / evidence synthesis agent,Complete,5,5,,4,false,,4,https://elicit.com/agent/ac8d5a74-e395-41f4-b435-7e13b4174c1a,Returned four verified eligible studies and one explicitly incomplete record with DOI links and limitations.
Elicit,2,2026-08-01,Free / evidence synthesis agent,Complete,4,4,,0,false,,4,https://elicit.com/agent/a371298a-4389-492f-864b-3b146ac4cb52,Repeat produced a DOI-linked eligibility table and retained null and long-delay boundary cases.
Elicit,3,2026-08-01,Free / evidence synthesis agent,Partial,7,7,,0,true,Monthly free allowance exhausted before final response completed,4,https://elicit.com/agent/aee48341-23e6-49d1-bc35-35ddb72a75d5,At least seven eligible-study rows were populated before the allowance barrier interrupted completion.
Scite,1,2026-08-01,Free / Assistant / GPT-5 nano,Partial,13,1,,0,true,Only one of 13 reported references exposed without upgrading,0.5,https://scite.ai/assistant/retrieval-practice-vs-rereading-in-higher-education-DxEaye,"The assistant reported 13 verified references, but the full answer and most sources were hidden."
Scite,2,2026-08-01,Free / Assistant / GPT-5 nano,Complete,2,2,,0,true,,0.3,https://scite.ai/assistant/retrieval-practice-vs-rereading-in-university-students-Yg90rx,The repeated prompt retrieved two unrelated oncology papers and no eligible education studies.
Scite,3,2026-08-01,Free / Assistant / GPT-5 nano,Blocked,,0,,0,true,Mandatory seven-day trial prompt,0.1,https://scite.ai/assistant,The third matched run was blocked before any result could be generated.
Consensus,1,2026-08-01,Free / Pro-message allowance,Complete,,,,5,false,,0.5,https://consensus.app/search/retrieval-practice-versus-rereading-retention/ez1kT9CIRiCG5VVAIsFj9A/,Structured answer with an eligible-study table; five sampled claims matched paper records and DOIs.
Consensus,2,2026-08-01,Free / Pro-message allowance,Complete,,,,0,false,,0.8,https://consensus.app/search/retrieval-practice-versus-rereading/KkSMNAObR9ysSOie9tof6A/,"Repeat again produced eligible-study extraction, a qualified conclusion and explicit disagreement."
Consensus,3,2026-08-01,Free / Pro-message allowance,Complete,,,,0,false,,0.8,https://consensus.app/search/retrieval-practice-versus-rereading/ZbQrY7CMTnW22dR5SsCz5Q/,Third run separated the dominant finding from conditional and restudy-favouring evidence.
Undermind,1,2026-08-01,Free / Deep Search,Complete,82,82,21,0,false,,0.8,https://app.undermind.ai/projects/e3db7c45-1a7e-4f98-b368-95bbd3b7bda5,Deep Search returned ranked papers and DOI links; the summary did not visibly surface disagreement.
Undermind,2,2026-08-01,Free / Deep Search,Complete,90,90,27,0,false,,1.4,,Second matched search completed and its CSV export was preserved locally.
Undermind,3,2026-08-01,Free / Deep Search,Complete,84,84,24,0,false,,1.4,,Third matched search completed and noted less-uniform evidence for some implementations.