BuilderBenchby the LAB
27 October 2026 · v3 · pack 3.0.0
4-hour build window · Extra High (xhigh) · Standard
full result · owner-finalized partial personal review
One owner-finalized partial personal review of one GPT-6.1 Terra and Kimi K3.1 Pro run on the v3.0.0 pack. This is not a universal model ranking or a statistical claim.
full result · BuilderBench test of 27 October 2026 · 2 AIs on this page
Codex harnesswon this test
48.88/100
API estimate in US dollars · partial usage · 16 of 18 records priced
API equivalent · not a bill
full result · BuilderBench test of 27 October 2026 · 2 AIs on this page
BuilderBench API harness (OpenRouter) harnessfinal score
46.68/100
API estimate in US dollars · partial usage · 15 of 18 records priced
API equivalent · not a bill
owner-finalized partial personal review. one run per setup. winning one test does not make an AI the best overall. 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.
swipe the table sideways to see every column.
| task | GPT-6.1 Terrascore /100 | Kimi K3.1 Proscore /100 |
|---|---|---|
| 01 · Writing10% of quality | 33.40 | 34 |
| 02 · Landing page10% of quality | 41 | 47 |
| 03 · App lead magnet20% of quality | 25 | 54 |
| 04 · Video editing10% of quality | 35 | 34.33 |
| 05 · Motion5% of quality | 29 | 39.50 |
| 06 · Thumbnails10% of quality | 46 | 50.80 |
| 07 · Browser + computer10% of quality | 62.50 | 12.50 |
| 08 · Research10% of quality | 66 | 53 |
| 09 · Automation10% of quality | 94 | 80 |
| 10 · Founder conversations5% of quality | 54.50 | 55 |
swipe the table sideways to see every column.
| task | GPT-6.1 TerraAPI-equivalent estimate | Kimi K3.1 ProAPI-equivalent estimate |
|---|---|---|
| Writing | US$0.21recorded usage | US$0.25recorded usage |
| Landing page | US$0.53recorded usage | US$1.01recorded usage |
| App lead magnet | US$1.19recorded usage | US$0.65partial usage |
| Video editing | US$1.25partial usage | US$3.01recorded usage |
| Motion | US$0.76recorded usage | US$0.48partial usage |
| Thumbnails | US$0.34partial usage | US$1.77recorded usage |
| Browser + computer | unavailableusage unavailable | unavailableusage unavailable |
| Research | US$0.23recorded usage | US$0.25recorded usage |
| Automation | US$0.24recorded usage | US$0.22partial usage |
| 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 |
|---|---|---|---|
| Standard | $1.00 | $0.05 | $5.00 |
Pricing page lists one Standard rate; no separate long-context rate shown. Cache writes: $1.25 (30m) · $2.50 (12h). Run on the Codex subscription, so this is a list-price comparison, not a bill.
US dollars · Standard processing · Official pricing: developers.openai.com · checked
| per 1M tokens | input | cache read | output |
|---|---|---|---|
| Standard | $2.00 | $0.20 | $10.00 |
1M context window; no separate long-context rate listed. Pinned to the first-party Moonshot AI standard endpoint; the highspeed endpoint (double price) is excluded.
US dollars · Standard processing · Official pricing: openrouter.ai · checked
API-equivalent estimates price recorded execution usage only. They are not bills. A setup without recorded usage has an unknown cost, not zero. Setup, grading, presentation and subscription fees are excluded. Actual incremental spend is unknown.
review conditions
Rated as Candidate A and Candidate B on the BuilderBench score deck; names were revealed after score lock.
who runs this: Lennox Saint. the v3 tasks come from real work for Threadify, the business he co-founded.
The owner rated the saved outputs and the recorded live lanes on the BuilderBench score deck. The bench played the owner in both founder conversations from saved context. Scores were locked before the reveal. Quality is the weighted task score.
frozen setup
exact model and harness version strings are not in this public export, so they stay unknown here.
limits
p.s.
live count, 28 oct 2026 / 29 oct 2026.