Search-SWE

DeepSeek V4.1 Flash 1017

reasoning effort max · Last Updated

deepseek/deepseek-v4.1-flash-1017

Task results

Task catalog ↗
Task results for DeepSeek V4.1 Flash 1017
TaskHardwareAgentProviderTrajectory
agentic-search CPUCodex (v0.164.0)OpenRouter0.02h 0min78,190,365130,91677,850,028$0.36Trace →
agentic-search-vectorless CPUCodex (v0.164.0)OpenRouter0.02h 0min13,714,12897,01110,971,392$0.50Trace →
biomedical-graph-retrieval CPUCodex (v0.164.0)OpenRouter0.01h 44min82,338,985321,57282,067,968$0.48Trace →
budgeted-bimetric-retrieval CPUCodex (v0.164.0)OpenRouter0.51h 7min23,005,761151,96521,580,032$0.37Trace →
code-context-retrieval CPUCodex (v0.164.0)OpenRouter16.71h 59min54,729,695211,43553,845,504$0.42Trace →
evolving-personal-memory CPUCodex (v0.164.0)OpenRouter19.541min56,260,315305,11954,305,408$0.64Trace →
filtered-ann-retrieval CPUCodex (v0.164.0)OpenRouter0.02h 0min31,691,383282,14430,673,024$0.41Trace →
financial-report-qa CPUCodex (v0.164.0)OpenRouter6.91h 41min65,787,581258,60064,804,352$0.50Trace →
icsi-meeting-memory CPUCodex (v0.164.0)OpenRouter11.12h 0min57,698,115322,62157,056,512$0.46Trace →
incremental-vector-search CPUCodex (v0.164.0)OpenRouter7.62h 0min37,882,260288,81834,764,126$0.75Trace →
join-aware-table-retrieval CPUCodex (v0.164.0)OpenRouter30.01h 2min17,658,078147,51816,539,934$0.31Trace →
large-scale-formula-retrieval CPUCodex (v0.164.0)OpenRouter0.01h 20min5,526,257107,0665,419,648$0.10Trace →
long-document-reranking CPUCodex (v0.164.0)OpenRouter18.02h 0min9,965,677108,5639,717,769$0.13Trace →
long-pdf-evidence CPUCodex (v0.164.0)OpenRouter16.72h 0min41,466,104204,49940,936,576$0.32Trace →
long-pdf-evidence-vectorless CPUCodex (v0.164.0)OpenRouter50.02h 0min16,763,261117,45116,569,344$0.15Trace →
memory-constrained-dense-retrieval CPUCodex (v0.164.0)OpenRouter0.01h 29min73,957,570326,16873,099,506$0.54Trace →
narrative-clue-search CPUCodex (v0.164.0)OpenRouter9.12h 0min29,003,400176,65026,874,624$0.51Trace →
negative-constraint-retrieval CPUCodex (v0.164.0)OpenRouter70.01h 11min30,450,951214,84229,446,016$0.37Trace →
paper-metadata-retrieval CPUCodex (v0.164.0)OpenRouter12.51h 45min32,263,145187,63231,792,320$0.28Trace →
reasoning-query-rewriting CPUCodex (v0.164.0)OpenRouter30.01h 58min59,851,572159,21958,909,132$0.41Trace →
rust-bm25-search CPUCodex (v0.164.0)OpenRouter0.038min25,640,463193,32323,552,768$0.50Trace →
scientific-paper-qa CPUCodex (v0.164.0)OpenRouter0.02h 0min80,020,322221,21979,240,576$0.49Trace →
sparse-retrieval-pruning CPUCodex (v0.164.0)OpenRouter0.01h 59min89,879,717326,59588,309,151$0.70Trace →
visual-document-qa CPUCodex (v0.164.0)OpenRouter21.71h 55min43,405,107375,54539,592,146$0.92Trace →
budgeted-candidate-reranking GPUCodex (v0.164.0)OpenRouter17.12h 0min60,927,610237,77660,500,864$0.39Trace →
chunqiu-temporal-retrieval GPUCodex (v0.164.0)OpenRouter0.41h 33min66,484,650228,77765,267,152$0.52Trace →
code-embedding-finetuning GPUCodex (v0.164.0)OpenRouter17.04h 58min74,271,355235,38972,477,442$0.63Trace →
code-reranker-finetuning GPUCodex (v0.164.0)OpenRouter0.038min3,470,08064,2283,200,384$0.09Trace →
composed-image-search GPUCodex (v0.164.0)OpenRouter29.62h 0min15,141,433202,64013,275,468$0.44Trace →
constrained-molecular-search GPUCodex (v0.164.0)OpenRouter13.64h 17min214,341,252306,148210,097,152$1.45Trace →
decision-model-training GPUCodex (v0.164.0)OpenRouter0.016min409,275147,003315,776$0.10Trace →
document-page-search GPUCodex (v0.164.0)OpenRouter0.02h 0min10,434,763133,4309,393,920$0.26Trace →
instruction-following-embedding GPUCodex (v0.164.0)OpenRouter89.31h 31min20,907,536156,62620,721,280$0.18Trace →
khmer-embedding-training GPUCodex (v0.164.0)OpenRouter20.95h 55min55,183,710230,63554,538,752$0.40Trace →
layout-page-search GPUCodex (v0.164.0)OpenRouter18.62h 14min18,318,879272,37417,764,608$0.30Trace →
legal-stance-retrieval GPUCodex (v0.164.0)OpenRouter30.24h 31min72,955,224239,27572,687,744$0.40Trace →
lightweight-query-retrieval GPUCodex (v0.164.0)OpenRouter0.04h 19min265,369,029404,644262,744,802$1.42Trace →
methodology-inspiration-retrieval GPUCodex (v0.164.0)OpenRouter1.44h 18min64,544,579208,14063,879,040$0.42Trace →
multilingual-vector-postprocessing GPUCodex (v0.164.0)OpenRouter100.02h 4min30,772,324205,32430,299,581$0.29Trace →
protein-text-bidirectional-retrieval GPUCodex (v0.164.0)OpenRouter2.51h 0min6,442,099172,7056,052,064$0.18Trace →
quantity-aware-retrieval GPUCodex (v0.164.0)OpenRouter0.012min1,943,45968,3241,843,712$0.06Trace →
query-encoder-alignment GPUCodex (v0.164.0)OpenRouter0.03h 22min166,139,205304,899165,714,320$0.74Trace →
same-place-image-search GPUCodex (v0.164.0)OpenRouter52.02h 0min37,915,500205,06637,245,891$0.34Trace →
sketch-text-search GPUCodex (v0.164.0)OpenRouter0.06h 0min18,601,297155,38918,204,346$0.21Trace →
spoken-query-search GPUCodex (v0.164.0)OpenRouter14.32h 0min45,669,931355,06544,942,065$0.46Trace →
training-error-tracing GPUCodex (v0.164.0)OpenRouter0.02h 51min103,296,442330,36599,769,088$1.03Trace →
video-moment-search GPUCodex (v0.164.0)OpenRouter0.0<1min65,8087,59233,920$0.01Trace →

47 selected evaluations · valid zero scores included

Score: task reward × 100. Tokens: coding usage; cached tokens are included in input.

Cost is estimated from recorded agent tokens at official API rates in USD (2026-11-01; DeepSeek off-peak pricing).