Skip to content
  1. Home
  2. Benchmarks
  3. Integration Bench
  4. Kimi K3

Kimi K3

Integration capability of coding agents on Kimi K3.

Rank #14 of 21 Pass 1 57.47 ±4.96 Resolved 58% of tasks Harness OpenCode Effort default Total cost $92.79 Mean time 35 min Mean tool calls 56 Graded 50/50
61
polling
62% of tasks touching this surface resolved
66
writeback
67% of tasks touching this surface resolved
17
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 store9229/3049 (3 failed)18 min$1.24resolved
task-0002Legal-hold export: TalentForge custodians and everything on file against them10064/6453 (3 failed)29 min$1.49resolved
task-0003Build the weekly Placemint contact sweep100343/34354 (3 failed)22 min$1.57resolved
task-0004Merge StaffLine application stage with Placemint placement stage049/6589 (5 failed)61 min$2.47mandatory fail timed out
task-0005Keep TalentForge candidates fresh: webhooks + reconciliation polling018/2179 (5 failed)34 min$3.35mandatory fail
task-0006The case-note mirror is short every morning100159/15955 (4 failed)63 min$1.70resolved timed out
task-0007Harden the Vettly sync connector for production scheduling10039/3963 (4 failed)36 min$2.46resolved
task-0008Build a StaffLine full-sync + writeback connector032/3364 (4 failed)38 min$2.17mandatory fail
task-0009Migrate off the legacy Bullpen app-token integration10040/4067 (2 failed)44 min$2.74resolved
task-0010Build the GlobalHire audit-period extract Compliance keeps asking us for10081/8147 (3 failed)14 min$1.13resolved
task-0011TalentForge connector: consume webhooks while pushing corrections upstream033/5162 (4 failed)14 min$3.81mandatory fail
task-0012Build the full TalentLoop connector: polling, webhooks, and writeback89119/12375 (1 failed)28 min$3.04resolved
task-0013The warehouse and RecruitOS have been out of step for months093/9650 (4 failed)25 min$1.12partial
task-0014HireWire: clear the stale-screening correction backlog, exactly once10016/1649 (1 failed)28 min$1.29resolved
task-0015Bulk-import a migration batch into StaffLine020/2569 (11 failed)29 min$3.68mandatory fail
task-0016HireWire connector: push stage-change events + keep an incremental poll fresh9242/4556 (3 failed)27 min$2.05resolved
task-0017Move the Brightmoor candidate mirror onto TalentForge's event subscription0127/16063 (3 failed)33 min$2.13partial
task-0018Rebuild the nightly requisition rollup on the RecruitOS API100142/14252 (4 failed)13 min$0.89resolved
task-0019Rebuild the Vettly change feed for the vaultd cutover100184/18455 (3 failed)17 min$1.50resolved
task-0020Sync GlobalHire candidates into our canonical store (offset polling)10034/3454 (4 failed)16 min$0.86resolved
task-0021Sync GlobalHire candidates, placements, and agencies into our canonical store07/2333 (2 failed)60 min$0.62mandatory fail timed out
task-0022Harden the GlobalHire backfill for a production tenant10028/2886 (6 failed)64 min$2.24resolved
task-0023TalentLoop connector: keep a candidate/application mirror deletion-accurate018/2644 (2 failed)32 min$1.30mandatory fail
task-0024TalentLoop connector: unified freshness across webhook-covered and poll-only entities028/4061 (2 failed)27 min$2.70mandatory fail
task-0025Reset the CrewCall roster watermark after the tenant rebuild0500/50346 (2 failed)15 min$0.86partial
task-0026Push the Q1 close back into Placemint0269/26955 (4 failed)9 min$2.67partial
task-0027Reliability: Interviewly connector loses and regresses updates013/2555 (2 failed)61 min$1.97mandatory fail timed out
task-0028Harden the CrewCall snapshot connector for production10044/4474 (6 failed)37 min$2.00resolved
task-0029Get the Riverside signup file onto the CrewCall roster100102/10244 (2 failed)17 min$1.19resolved
task-0030Census the desk board against the Placemint placement book100325/32548 (4 failed)32 min$4.05resolved
task-0031Rebuild the morning bookable-crew board100101/10148 (3 failed)60 min$1.74resolved timed out
task-0032Build the weekly capability index036/1171460 min$0.06partial timed out
task-0033Keep the RecruitOS pipeline and the Placemint ledger in step100153/15360 (6 failed)53 min$2.39resolved
task-0034Weekly crew top-up is putting people on the gate twice100165/16574 (4 failed)20 min$1.82resolved
task-0035Sync Vettly subjects, checks, and reports into our canonical store10028/2851 (4 failed)16 min$1.61resolved
task-0036Harden the Placemint book extract for the larger tenant0212/42032 (2 failed)60 min$0.87partial timed out
task-0037Fix the Placemint redeployment sync100306/30679 (6 failed)39 min$2.78resolved
task-0038TalentLoop connector: make webhook-driven freshness durable against a misbehaving receiver018/3740 (2 failed)62 min$1.26mandatory fail timed out
task-0039Build the nightly payroll bridge for the punch export10094/9458 (4 failed)21 min$2.09resolved
task-0040Push interview reschedules upstream into Interviewly018/2080 (4 failed)26 min$2.09mandatory fail
task-0041Paygrade connector: employee/assignment sync and writeback030/3854 (2 failed)40 min$1.64mandatory fail
task-0042Move the rota mirror onto instant storage100117/11745 (5 failed)11 min$1.11resolved
task-0043The rota warehouse is counting the same night more than once083/8927 (3 failed)60 min$0.54partial timed out
task-0044Nightly roster census disagrees with CrewCall about who works here10067/6743 (2 failed)9 min$0.67resolved
task-0045Replatform the ATS integration: StaffLine to Bullpen v2020/2471 (5 failed)47 min$4.04mandatory fail
task-0046Get the scorecard export into TalentLoop10088/8846 (1 failed)35 min$1.80resolved
task-0047Reconcile and safely apply the mobility stage repair queue098/10072 (2 failed)44 min$2.11partial
task-0048Migrate the Brackett closure archive onto Paygrade100211/21144 (3 failed)50 min$1.44resolved
task-0049Placemint connector: webhooks, polling, and writeback for a high-volume tenant10065/6548 (5 failed)68 min$1.61resolved
task-0050Fix the Placemint fee-corrections connector100254/25443 (6 failed)10 min$0.80resolved