BuilderBenchby the LAB
14 October 2026 · v3 · pack 3.0.0
4-hour build window · Extra High requested (Argon ran on High) · Standard
the reveal · owner-finalized partial personal review
One owner-finalized partial personal review of one Gemini 4 Argon run through the BuilderBench API harness and one Muse Wonder Full run in Meta's Muse Code app on the v3 pack. This is not a universal model ranking or a statistical claim.
BuilderBench API harnessleads this assessment
45.38/100
API estimate in US dollars · partial usage · 15 of 31 records priced
API equivalent · not a bill
Muse Code harnessfinal score
39.91/100
API estimate in US dollars · usage unavailable · 0 of 23 records priced
API equivalent · not a bill
owner-finalized partial personal review. one run per setup. a lead in one assessment is not a ranking. the limits are below.
quality is the weighted average of the task scores. autonomy is how little owner rescue the setup needed. proactivity is useful initiative inside the brief. each is scored out of 100.
the final score is quality × 70% + autonomy × 20% + proactivity × 10%. cost never buys points.
dated references
each bar is one owner review of one run. the packs, tasks and evidence limits differ, and each bar says which pack and which kind of review it was. scores only compare inside one pack.
task breakdown
each task is scored out of 100. its weight is its share of the quality score.
| task | Gemini 4 Argonscore /100 | Muse Wonder Fullscore /100 |
|---|---|---|
| 01 · Writing10% of quality | 50 | 40 |
| 02 · Landing page10% of quality | 35 | 60 |
| 03 · App lead magnet20% of quality | 20 | 20 |
| 04 · Video editing10% of quality | 0 | 33.33 |
| 05 · Motion5% of quality | 25 | 30 |
| 06 · Thumbnails10% of quality | 62 | 39 |
| 07 · Browser + computer10% of quality | 20 | 0 |
| 08 · Research10% of quality | 45 | 50 |
| 09 · Automation10% of quality | 91.67 | 75 |
| 10 · Founder conversations5% of quality | 70 | 50 |
| task | Gemini 4 ArgonAPI-equivalent estimate | Muse Wonder FullAPI-equivalent estimate |
|---|---|---|
| Writing | US$1.05partial usage | unavailableusage unavailable |
| Landing page | US$1.00partial usage | unavailableusage unavailable |
| App lead magnet | US$2.73partial usage | unavailableusage unavailable |
| Video editing | unavailableusage unavailable | unavailableusage unavailable |
| Motion | unavailableusage unavailable | unavailableusage unavailable |
| Thumbnails | US$3.00partial usage | unavailableusage unavailable |
| Browser + computer | unavailableusage unavailable | unavailableusage unavailable |
| Research | US$0.40partial usage | unavailableusage unavailable |
| Automation | US$0.74partial usage | unavailableusage unavailable |
| Founder conversations | unavailableusage unavailable | unavailableusage unavailable |
estimates, not bills. “unavailable” means no usage was recorded for that task. it does not mean the task was free.
native sessions
these four sessions run in the real app, outside the build window, and each gets its own rating. browser and computer average into the browser + computer task. advice and planning average into the founder conversations task. see the four prompts
API pricing
the rates are what the provider listed on the date shown. the estimate prices the usage this run recorded. the bill is what was actually charged. they are kept apart on purpose.
| per 1M tokens | input | cache read | output |
|---|---|---|---|
| Introductory, through 30 Nov 2026 | $2.00 | $0.10 | $10.00 |
From 1 December 2026: $4.00 input, $0.20 cache read and $20.00 output. Output includes thinking tokens. Cache reads are implicit; explicit cache writes are unsupported. 1M context with no separate long-context price. API use is billed separately from Google subscriptions.
US dollars · Standard processing · Google Gemini API docs: Gemini 4 Argon · checked
official API price not published.Meta offers Muse Wonder Full only inside Muse Code subscriptions (Power plan, $50 a month) and publishes no API token prices. A plan price is not a run bill, and regular Muse Wonder rates are not substituted. last checked 13 Oct 2026.no other model’s rate was used in its place.
API-equivalent estimates price recorded execution usage only. They are not bills. Gemini 4 Argon keeps provider-reported estimates for 15 of 31 saved records, so its figure is partial. Muse Wonder Full ran on Meta's Muse Code Power plan, which reported no usage for these runs, so its run cost is unknown, not zero. The $50-a-month plan price is not a run bill. Native sessions, setup, calibration, coordination, grading, presentation and subscription fees are excluded. Actual incremental spend is unknown.
review conditions
Requested as a blind comparison. The main build jobs were rated as Candidate A and Candidate B, and the names were revealed after score lock. It was not fully blind: operator logs leaked identity clues, autonomy and proactivity were rated by name, and the native sessions were rated by model name in a separate form.
The owner rated the saved build outputs, the four native sessions, autonomy and proactivity. The automation repair task was conductor-scored from recorded evidence. The saved scores were finalized as a partial personal review and locked before the names were revealed. Quality is the weighted task score. v1 to v3 use 70% quality, 20% autonomy and 10% proactivity.
frozen setup
exact model and harness version strings are not in this public export, so they stay unknown here.
limits
p.s.
live count, 16 oct 2026 / 17 oct 2026.