FAVUR

2048 - the sliding-tile game

Run of , staffed by a mixed roster (google + meta). Combines Muse Spark (primary) with minority vendors for cost savings - composite 59.58.

The verdict.

58.94 composite · rubric v15

duration
218 min
requests
2570
failed requests
13 (0.51%)
cost
$99.16

Scores come from the open scoring pipeline - methodology and leaderboards at Favur Evals ↗

Per-subject scores.

subjectscore
code quality 32.42
cost efficiency 33.48
deliverables 36.6
effort efficiency 44.49
process discipline 48.44
test quality 48.63
tool discipline 83.63
velocity 52.13

Who staffed it.

rolemodel
codemeta/muse-spark-1.2
specmeta/muse-spark-1.2
testgoogle/gemini-3.5-flash-lite
buildgoogle/gemini-3.5-flash-lite
scoutmeta/muse-spark-1.2
developgoogle/gemini-3.5-flash-lite
platformmeta/muse-spark-1.2
architectmeta/muse-spark-1.2
conciergegoogle/gemini-3.5-flash-lite
pseudocodemeta/muse-spark-1.2
code-reviewmeta/muse-spark-1.2
sprint-planmeta/muse-spark-1.2
orchestratorgoogle/gemini-3.5-flash-lite
sprint-reviewmeta/muse-spark-1.2

What shipped.

tests
57/93 passing
source
1533 lines
test code
978 lines
coverage
46.15%

Detected moments.

  • At peak, 5 agents were mid-tool-call in the same second. One run, one job - and a headcount no single-agent tool can log. 234 agents were hired across 4 hours; 3 different models split the work by role. A dedicated supervision lane spent 375 model calls doing nothing but checkups on the rest of the team.
  • One agent generated 13 images in 2 minutes. An agent can call an image model mid-run when the job needs art it cannot draw itself. The spec agent produced 13 generated images in 2 minutes, one after another. The set landed in the project as files - a coherent batch of assets built to order in one sitting.
  • The concierge argued a product call for 1 round - and locked a rule. A concierge agent wanders the repo full-time, second-guessing product quality nobody asked it to check. It opened a debate and worked through pros, cons and risks over 1 round, pushing back on its own first answer. The debate ended in a locked rule the rest of the team now builds under.
  • The concierge argued a product call for 1 round - and locked a rule. A concierge agent wanders the repo full-time, second-guessing product quality nobody asked it to check. It opened a debate and worked through pros, cons and risks over 1 round, pushing back on its own first answer. The debate ended in a locked rule the rest of the team now builds under.
  • The concierge argued a product call for 1 round - and locked a rule. A concierge agent wanders the repo full-time, second-guessing product quality nobody asked it to check. It opened a debate and worked through pros, cons and risks over 1 round, pushing back on its own first answer. The debate ended in a locked rule the rest of the team now builds under.
  • The same command failed 4 times in 7 minutes - then passed. Agents run their own build and test commands, and in a team of agents the code can change under a command between runs. One agent ran the same command, `poetry run ruff format --check .`, 4 times over 7 minutes; every run failed. The command never changed - the code underneath it did: 11 file writes landed inside the loop window, and the next run passed.
  • The run's priciest five minutes - 2.7 times a typical window. Every model call is metered, so a run's spend can be read window by window against its own baseline. For five minutes, 118 model requests from 11 agents landed at once - 2.7 times the run's typical five-minute spend. The burst bought something: the checkup agent finished its job.
  • Sprint 1 came back rejected - the rework passed. At the end of each sprint, a sprint-review agent judges the whole sprint's output at once. It sent sprint 1 back: single HIGH severity test-product separation finding (_Clock wrapper) blocks APPROVE. The team redid the work and the next review signed off; the sprint closed with 5 of its 7 planned tasks complete.