Skip to content
  1. Home
  2. Benchmarks
  3. Integration Bench
  4. DeepSeek V4 Flash (max)

DeepSeek V4 Flash (max)

Integration capability of coding agents on DeepSeek V4 Flash (max).

Rank #16 of 21 Pass 1 55.68 ±4.68 Resolved 56% of tasks Harness OpenCode Effort max Total cost $4.28 Mean time 24 min Mean tool calls 82 Graded 50/50
62
polling
62% of tasks touching this surface resolved
57
writeback
57% of tasks touching this surface resolved
18
webhooks
18% of tasks touching this surface resolved

Published attempts

50 of 50 scored tasks released

One official run per task. Click a row for the full trajectory: every tool call, vendor request, graded check and the diff.

TaskTicketScoreChecksTool callsTimeCostOutcome
task-0001Sync StaffLine candidates, jobs, applications, and notes into our canonical store9229/3093 (3 failed)26 min$0.08resolved
task-0002Legal-hold export: TalentForge custodians and everything on file against them10064/6445 (4 failed)11 min$0.03resolved
task-0003Build the weekly Placemint contact sweep100343/34380 (5 failed)24 min$0.08resolved
task-0004Merge StaffLine application stage with Placemint placement stage048/6493 (4 failed)23 min$0.09mandatory fail
task-0005Keep TalentForge candidates fresh: webhooks + reconciliation polling017/21151 (12 failed)51 min$0.25mandatory fail
task-0006The case-note mirror is short every morning100159/15983 (6 failed)43 min$0.07resolved
task-0007Harden the Vettly sync connector for production scheduling10039/3977 (6 failed)38 min$0.06resolved
task-0008Build a StaffLine full-sync + writeback connector10033/33100 (2 failed)20 min$0.08resolved
task-0009Migrate off the legacy Bullpen app-token integration10040/4092 (6 failed)9 min$0.09resolved
task-0010Build the GlobalHire audit-period extract Compliance keeps asking us for10081/8169 (2 failed)10 min$0.04resolved
task-0011TalentForge connector: consume webhooks while pushing corrections upstream033/51149 (7 failed)52 min$0.23mandatory fail
task-0012Build the full TalentLoop connector: polling, webhooks, and writeback00/083 (4 failed)31 min$0.07partial
task-0013The warehouse and RecruitOS have been out of step for months10096/9646 (6 failed)18 min$0.04resolved
task-0014HireWire: clear the stale-screening correction backlog, exactly once06/1643 (2 failed)15 min$0.02mandatory fail
task-0015Bulk-import a migration batch into StaffLine020/2580 (3 failed)6 min$0.08mandatory fail
task-0016HireWire connector: push stage-change events + keep an incremental poll fresh9245/48107 (4 failed)33 min$0.11resolved
task-0017Move the Brightmoor candidate mirror onto TalentForge's event subscription0143/16060 (3 failed)26 min$0.04partial
task-0018Rebuild the nightly requisition rollup on the RecruitOS API100142/14252 (4 failed)11 min$0.02resolved
task-0019Rebuild the Vettly change feed for the vaultd cutover100183/18455 (3 failed)18 min$0.05resolved
task-0020Sync GlobalHire candidates into our canonical store (offset polling)10034/3489 (5 failed)8 min$0.08resolved
task-0021Sync GlobalHire candidates, placements, and agencies into our canonical store015/2370 (5 failed)14 min$0.04mandatory fail
task-0022Harden the GlobalHire backfill for a production tenant10028/2893 (7 failed)59 min$0.10resolved
task-0023TalentLoop connector: keep a candidate/application mirror deletion-accurate018/26105 (7 failed)38 min$0.11mandatory fail
task-0024TalentLoop connector: unified freshness across webhook-covered and poll-only entities031/40102 (6 failed)13 min$0.14mandatory fail
task-0025Reset the CrewCall roster watermark after the tenant rebuild0500/50358 (4 failed)9 min$0.04partial
task-0026Push the Q1 close back into Placemint100269/26989 (4 failed)33 min$0.11resolved
task-0027Reliability: Interviewly connector loses and regresses updates012/2449 (3 failed)20 min$0.04mandatory fail
task-0028Harden the CrewCall snapshot connector for production10044/44108 (8 failed)36 min$0.11resolved
task-0029Get the Riverside signup file onto the CrewCall roster060/10276 (5 failed)6 min$0.08partial
task-0030Census the desk board against the Placemint placement book100325/32564 (8 failed)5 min$0.05resolved
task-0031Rebuild the morning bookable-crew board100101/10170 (5 failed)19 min$0.05resolved
task-0032Build the weekly capability index100118/11856 (5 failed)13 min$0.05resolved
task-0033Keep the RecruitOS pipeline and the Placemint ledger in step100153/153108 (8 failed)42 min$0.15resolved
task-0034Weekly crew top-up is putting people on the gate twice100165/165101 (6 failed)27 min$0.12resolved
task-0035Sync Vettly subjects, checks, and reports into our canonical store025/2881 (6 failed)16 min$0.09mandatory fail
task-0036Harden the Placemint book extract for the larger tenant0212/42055 (6 failed)13 min$0.05partial
task-0037Fix the Placemint redeployment sync100306/306130 (9 failed)52 min$0.23resolved
task-0038TalentLoop connector: make webhook-driven freshness durable against a misbehaving receiver021/3756 (2 failed)23 min$0.04mandatory fail
task-0039Build the nightly payroll bridge for the punch export10094/9462 (6 failed)4 min$0.05resolved
task-0040Push interview reschedules upstream into Interviewly015/17119 (9 failed)35 min$0.17mandatory fail
task-0041Paygrade connector: employee/assignment sync and writeback030/38102 (4 failed)59 min$0.16mandatory fail
task-0042Move the rota mirror onto instant storage100117/11764 (5 failed)26 min$0.08resolved
task-0043The rota warehouse is counting the same night more than once083/8955 (2 failed)5 min$0.05partial
task-0044Nightly roster census disagrees with CrewCall about who works here10067/6765 (8 failed)6 min$0.05resolved
task-0045Replatform the ATS integration: StaffLine to Bullpen v2021/25110 (8 failed)8 min$0.12mandatory fail
task-0046Get the scorecard export into TalentLoop084/8951 (3 failed)29 min$0.07partial
task-0047Reconcile and safely apply the mobility stage repair queue099/1009025 min$0.08partial
task-0048Migrate the Brackett closure archive onto Paygrade100211/21168 (2 failed)25 min$0.08resolved
task-0049Placemint connector: webhooks, polling, and writeback for a high-volume tenant10068/68106 (7 failed)44 min$0.13resolved
task-0050Fix the Placemint fee-corrections connector100254/25463 (2 failed)15 min$0.05resolved