← older: fast__qwen3.6-35b-a3b__20260708-210920all runs · task boardnewer: smoke__qwen3.6-35b-a3b__20260708-233720

fast__qwen3.6-27b__20260708-213445

QwenPi (qwen-native tools) — 27b fast, 5/9; format-leak still bites regex-log

QwenPi harness, 27b, fast set (loop_guard on, strip off, subagent_timeout 600; sparql cut@4m by the fast cap but PASSED).

Result 5/9 (fix-git, nginx, sparql, sanitize-git-repo, openssl PASS; query-optimize, regex-log, filter-js, cancel-async FAIL). Lower than the old plain-MinimalPi 27b baseline (…20260706-124744, 7/9) but that baseline is pre-many-harness-changes and the two swing tasks (query-optimize, regex-log) are known-flaky for 27b (the 07-06 K=2 run …132725 split them 1/1). Not a clean score delta.

Tool mechanism verdict — clean: 0 schema/unknown-tool errors; qwen-named tools used fluently.

KEY SHAPE FINDING — regex-log FAIL is the wire-format leak, and qwen_tools does NOT fix it (as predicted): the model emitted its next tool call as raw qwen XML inside the THINKING block — <function=run_shell_command><parameter=command>which python… — note it now uses the qwen NAME run_shell_command, but the call sits in the reasoning channel so llama.cpp never parses it → empty final message → task untouched (exact 56Avdhy shape). This is a template/parser bug, orthogonal to tool naming; the fix is the empty-final-message recovery ledger item (or a llama.cpp qwen-XML/thinking fix), not the tool surface.

Run details

modelllama-local/qwen3.6-27bagentharnesses.qwen_pi:QwenPithinkingonreasoning budgetdirect (server/none)* — see journal for the authoritative mechanismmaxTokens / contextWindow65536 / 196608agent timeout ×2.0trials9 of 9 — 5 pass · 4 failmean reward0.56tokens (job total)1,081,762 in / 52,854 outstarted / finished2026-07-08T21:34 / 2026-07-08T22:13wall clock38m50s

Tasks

cancel-async-tasks — 0/1 passed

#resulttotalagentin/out tokflags
1FAIL3m21s2m29s60905/6773
a bash command timed out
🔍 view

filter-js-from-html — 0/1 passed

#resulttotalagentin/out tokflags
1FAIL6m56s3m15s178520/10074
🔍 view

fix-git — 1/1 passed

#resulttotalagentin/out tokflags
1PASS1m14s35s34643/1574
🔍 view

nginx-request-logging — 1/1 passed

#resulttotalagentin/out tokflags
1PASS1m11s33s29887/1650
🔍 view

openssl-selfsigned-cert — 1/1 passed

#resulttotalagentin/out tokflags
1PASS1m34s58s38456/3195
🔍 view

query-optimize — 0/1 passed

#resulttotalagentin/out tokflags
1FAIL12m49s6m12s97437/4518
a bash command timed out
🔍 view

regex-log — 0/1 passed

#resulttotalagentin/out tokflags
1FAIL3m58s3m10s57214/10246
empty final message (no text, no tool call)long reasoning (19,782 chars)
🔍 view

sanitize-git-repo — 1/1 passed

#resulttotalagentin/out tokflags
1PASS2m56s2m17s379481/4718
🔍 view

sparql-university — 1/1 passed

#resulttotalagentin/out tokflags
1PASS4m48s4m00s205219/10106
fast-timeout cut at 4mlong reasoning (25,006 chars)
🔍 view