From fe73112aa2c7f4bb447a592ed8e068c71e98b000 Mon Sep 17 00:00:00 2001 From: Don Garrett Date: Tue, 29 Sep 2026 14:18:50 +0000 Subject: [PATCH] Use qwen3.8-27b as the example executor It led the executor benchmark (PR #55): 37/40 over eight tasks and 17/20 on the harder four, ahead of claude-haiku-4-5's 15/20. local-fast, the previous example, scored 0/8 on the harder four. Model selection is on the dev executor alone for now; the shared-host question (spec Part H) stays open. Co-Authored-By: Claude Opus 5.5 (1M context) --- nisse-dev.example.yaml | 2 +- nisse-dev/tests/test_config.py | 4 +++- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/nisse-dev.example.yaml b/nisse-dev.example.yaml index 6bc70ce..c8b9c11 100644 --- a/nisse-dev.example.yaml +++ b/nisse-dev.example.yaml @@ -23,7 +23,7 @@ llm: # A5: LiteLLM is the only LLM destination request_timeout_s: 600 roles: # A11: role -> LiteLLM model alias - executor: local-fast # the local tier; the benchmark picks the model + executor: qwen3.8-27b # the local tier: best executor in the benchmark (PR #55) planner: claude-opus-5-5 # high-end judgement: plans a weaker executor follows delegate: claude-opus-5-5 # takes over what the executor got stuck on consultant: claude-haiku-4-5 # one question at a time, no tools: cheap is enough diff --git a/nisse-dev/tests/test_config.py b/nisse-dev/tests/test_config.py index e5c7bcb..eeb710a 100644 --- a/nisse-dev/tests/test_config.py +++ b/nisse-dev/tests/test_config.py @@ -272,7 +272,9 @@ def test_the_example_file_is_valid() -> None: "NISSE_MATRIX_TOKEN": "syt_x", "NISSE_GITEA_TOKEN": "gt", } - assert AgentConfig.from_file(example, env).tuning.executor.max_turns == 40 + config = AgentConfig.from_file(example, env) + assert config.tuning.executor.max_turns == 40 + assert config.roles.executor == "qwen3.8-27b" # --- nothing that carries secrets may sit inside the clone ----------------- -- 2.54.0