Batch 2 took about half the time and one-third the modeled spend. Batch 1 included all 20 branch histories with either approach.
Batch 2 observed 49% less time and 67% lower modeled spend with Merget (57% lower at list price). It incorporated 23 branch histories with 15 direct merges and held one for review.
How to read thisMedians of 3–4 completed runs; thin lines show the observed range.
Batch 2 · flask · 24 branches6 conflicted merges in the original history, all within one group of related branches; replaying it, the agents met 7 (one the replay itself produced) and validated all 7
Batch 1 included all 20 branch histories with either approach. Merget’s median time was 13% higher and modeled spend 19% higher (82% higher at list price). Fewer direct merges did not mean less branch history included.
How to read thisMedians of 3–4 completed runs; thin lines show the observed range.
Batch 1 · flask · 20 branches8 conflicted merges in the original history, all within two groups of related branches
The one-call resolver cut modeled spend by 69% on the same batch outcomes. It reduced agent sessions from 7 to 2, with 26% less observed time.
Flask batch 2, medians of 3–4 completed runs per configuration.
No new test failures in any completed run. Merget also preserved Git’s already-merged text in every auditable merge.
- Clean results. No finished run left a conflict marker, an open Markdown fence, an edit outside the conflicts or a new test failure at the end.
- Edits outside the conflict. In the audited merges, sequential agents altered git’s merged text once in each run of batch 1; Merget altered none.
A branch can arrive through its own merge or inside another branch’s history. Both count as work delivered.
A branch gets its own merge commit. A later branch can include earlier branch history without another direct merge.
Contained branches enter through another branch’s history; carried-in branches arrive with a later successful merge. “Validated” counts direct conflicted merges that pass the test recipe.
Merget’s first way to resolve a conflict: one model call per conflicted file, with no agent. When it is off or fails, Merget falls back to an agent session.
Modeled spend charges cached tokens at the full input rate. List-price spend charges cached tokens at their discounted rate. We quote both.
Every completed run, with each unit named. Direct merges, included history, checks and resource use.
Full results: medians on every measure
Batch 1 included all 20 branch histories with either approach. Merget’s median time was 13% higher and modeled spend 19% higher (82% higher at list price). Fewer direct merges did not mean less branch history included.
How to read thisMedians of 3–4 completed runs; thin lines show the observed range.
Batch 1 · flask · 20 branches8 conflicted merges in the original history, all within two groups of related branches
Batch 2 observed 49% less time and 67% lower modeled spend with Merget (57% lower at list price). It incorporated 23 branch histories with 15 direct merges and held one for review.
How to read thisMedians of 3–4 completed runs; thin lines show the observed range.
Batch 2 · flask · 24 branches6 conflicted merges in the original history, all within one group of related branches; replaying it, the agents met 7 (one the replay itself produced) and validated all 7
Full results: wall time of every run
One Merget run on batch 1 took 838.3 s, against about 1,890 s for the other three.
How to read thisEach square is one run’s wall time; runs with nearly the same time are stacked. The vertical tick marks the median.
Full results: every finished run
Every completed run: direct merges, branch history that arrived inside later merges, and checks.
| Batch | Approach | Run | Wall time | Direct merge operations | Arrived inside later merges | Carried in | Direct conflicts validated | New test failures |
|---|---|---|---|---|---|---|---|---|
| Batch 1 | Sequential agents | Run 1 | 1,674.0 s | 17 direct | 0 | 3 | 7 | 0 |
| Batch 1 | Sequential agents | Run 2 | 1,512.4 s | 17 direct | 0 | 3 | 7 | 0 |
| Batch 1 | Sequential agents | Run 3 | 1,684.6 s | 17 direct | 0 | 3 | 7 | 0 |
| Batch 1 | Merget | Run 1 | 1,911.3 s | 14 direct | 2 | 4 | 5 | 0 |
| Batch 1 | Merget | Run 1, repeated | 838.3 s | 14 direct | 2 | 4 | 5 | 0 |
| Batch 1 | Merget | Run 2 | 1,890.2 s | 14 direct | 2 | 4 | 5 | 0 |
| Batch 1 | Merget | Run 3 | 1,891.8 s | 14 direct | 2 | 4 | 5 | 0 |
| Batch 2 | Sequential agents | Run 1 | 709.1 s | 24 direct | 0 | 0 | 7 | 0 |
| Batch 2 | Sequential agents | Run 2 | 692.5 s | 24 direct | 0 | 0 | 7 | 0 |
| Batch 2 | Sequential agents | Run 3 | 581.6 s | 24 direct | 0 | 0 | 7 | 0 |
| Batch 2 | Merget | Run 1 | 354.0 s | 15 direct | 8 | 0 | 2 | 0 |
| Batch 2 | Merget | Run 1, repeated | 354.5 s | 15 direct | 8 | 0 | 2 | 0 |
| Batch 2 | Merget | Run 2 | 354.7 s | 15 direct | 8 | 0 | 2 | 0 |
| Batch 2 | Merget | Run 3 | 302.9 s | 15 direct | 8 | 0 | 2 | 0 |
| Batch 2 | Merget without the resolver | Run 1 | 484.0 s | 15 direct | 8 | 0 | 2 | 0 |
| Batch 2 | Merget without the resolver | Run 1, repeated | 410.1 s | 15 direct | 8 | 0 | 2 | 0 |
| Batch 2 | Merget without the resolver | Run 2 | 478.3 s | 15 direct | 8 | 0 | 2 | 0 |
The same runs: agent sessions, model calls and spend in both units.
| Batch | Approach | Run | Agent sessions | Resolver calls | Agent model calls | Modeled spend | List-price spend |
|---|---|---|---|---|---|---|---|
| Batch 1 | Sequential agents | Run 1 | 10 | 0 | 75 | $8.6558 | $2.5236 |
| Batch 1 | Sequential agents | Run 2 | 10 | 0 | 92 | $9.7428 | $2.6381 |
| Batch 1 | Sequential agents | Run 3 | 10 | 0 | 91 | $12.8790 | $2.9787 |
| Batch 1 | Merget | Run 1 | 17 | 76 | 149 | $11.3749 | $4.9979 |
| Batch 1 | Merget | Run 1, repeated | 16 | 77 | 134 | $9.6123 | $3.5246 |
| Batch 1 | Merget | Run 2 | 17 | 76 | 146 | $11.7355 | $4.5957 |
| Batch 1 | Merget | Run 3 | 16 | 77 | 148 | $12.7238 | $5.3383 |
| Batch 2 | Sequential agents | Run 1 | 7 | 0 | 54 | $4.8957 | $1.3978 |
| Batch 2 | Sequential agents | Run 2 | 7 | 0 | 46 | $2.9831 | $1.0561 |
| Batch 2 | Sequential agents | Run 3 | 7 | 0 | 49 | $3.2830 | $1.0243 |
| Batch 2 | Merget | Run 1 | 2 | 7 | 12 | $1.0231 | $0.5048 |
| Batch 2 | Merget | Run 1, repeated | 2 | 7 | 14 | $1.1477 | $0.4263 |
| Batch 2 | Merget | Run 2 | 2 | 7 | 12 | $1.0489 | $0.4346 |
| Batch 2 | Merget | Run 3 | 2 | 7 | 13 | $1.3223 | $0.4764 |
| Batch 2 | Merget without the resolver | Run 1 | 7 | 0 | 52 | $3.8563 | $1.1162 |
| Batch 2 | Merget without the resolver | Run 1, repeated | 7 | 0 | 49 | $3.1382 | $0.9619 |
| Batch 2 | Merget without the resolver | Run 2 | 7 | 0 | 48 | $3.6001 | $1.0576 |
Full results: what one resolver call and one agent session cost
What one resolver call and one agent session cost, modeled.
Use Merget to reduce repeated merge work. The strongest comparison isolates the resolver on the same branch outcomes.
The one-call resolver handled enough of batch 2 to replace five of seven full agent sessions. With the same included branch histories, median modeled spend fell from $3.60 to $1.10 and time fell from 478.3 to 354.3 seconds.
Against sequential agents, branch containment avoided repeated direct merges: on batch 2 Merget brought in 23 branch histories with 15 merges.
Cut the repeated merge work. Connect one repository and let Merget take the rebases your agents repeat.
Sign up freeMeasured November 1, 2026 on historical Flask branches: the same model and per-conflict budget for both approaches, 3 to 4 runs each.
Merget vs graphify: more complete call graphs