Python → Go
The implementation language moves, the artifact does not. 4,393 lines, 15 hours per model, and 8,530 checks across 15 modules.
- Runs
- 26
- Past rung 1
- 14
- Past rung 2
- 0
- Past rung 3
- 0
- Best composite
- —
one per configuration
of 26 — the migration happened
of those 14 — every check perfect
no run got this far
nothing scored here
The instrument
Acceptance gates
8
decide whether the migration happened
Behavioral modules
15
ported suites, run against the new tree
Frozen checks
8,530
all must pass to open rung three
Verifiers
6
hostile programs, written against the spec
Where the 26 runs stopped
- 12
- Never migrated
- 14
- Migrated, behavior lost
- 0
- Perfect suite, gate rejected
- 0
- Verifier found a difference
- 0
- Accepted
No run was perfect on all 8,530 checks. No run reached rung three, so no verifier was paid to attack this task.
By model
| Model | Runs | Gate | Ceil. | Where they stopped | Mean | Best |
|---|---|---|---|---|---|---|
| claude-opus-5 | 5 | 4 | 0 | | 0.0 | — |
| claude-sonnet-5 | 5 | 1 | 0 | | 0.0 | — |
| dsv4-flash | 1 | 0 | 0 | | 0.0 | — |
| glm-5.2 | 1 | 0 | 0 | | 0.0 | — |
| gpt-5.6-luna | 6 | 4 | 0 | | 0.0 | — |
| gpt-5.6-sol | 6 | 4 | 0 | | 0.0 | — |
| kimi-k3 | 1 | 1 | 0 | | 0.0 | — |
| qwen3.8-max | 1 | 0 | 0 | | 0.0 | — |
Every run
best composite firstOne row per configuration. Shape is the run's tool sequence in 24 slices — pale is shell, dark is an edit — and each log is the full session as recorded.
| Model | Effort | Outcome | Gate | Checks | Verif. | Score | Shape |
|---|---|---|---|---|---|---|---|
| claude-opus-5 | xhigh | partial | 8/8 | 97.4% | — | 0 | |
| claude-opus-5 | low | partial | 8/8 | 96.9% | — | 0 | |
| claude-opus-5 | medium | partial | 8/8 | 96.1% | — | 0 | |
| claude-opus-5 | high | partial | 8/8 | 95.3% | — | 0 | |
| claude-sonnet-5 | high | failed | 6/8 | 94.3% | — | 0 | |
| claude-opus-5 | max | failed | 7/8 | 88.7% | — | 0 | |
| gpt-5.6-sol | max | partial | 8/8 | 85.8% | — | 0 | |
| gpt-5.6-luna | max | partial | 8/8 | 84.9% | — | 0 | |
| claude-sonnet-5 | max | partial | 8/8 | 84.9% | — | 0 | |
| claude-sonnet-5 | low | failed | 6/8 | 84.3% | — | 0 | |
| qwen3.8-max | max | failed | 7/8 | 83.0% | — | 0 | |
| claude-sonnet-5 | medium | failed | 7/8 | 82.6% | — | 0 | |
| gpt-5.6-luna | xhigh | partial | 8/8 | 79.5% | — | 0 | |
| kimi-k3 | max | partial | 8/8 | 78.0% | — | 0 | |
| dsv4-flash | max | failed | 7/8 | 76.9% | — | 0 | |
| glm-5.2 | max | failed | 7/8 | 76.0% | — | 0 | |
| gpt-5.6-sol | medium | partial | 8/8 | 74.3% | — | 0 | |
| gpt-5.6-luna | high | partial | 8/8 | 72.5% | — | 0 | |
| gpt-5.6-sol | low | partial | 8/8 | 66.5% | — | 0 | |
| gpt-5.6-sol | xhigh | partial | 8/8 | 55.5% | — | 0 | |
| gpt-5.6-luna | medium | partial | 8/8 | 53.1% | — | 0 | |
| gpt-5.6-luna | none | failed | 7/8 | 39.0% | — | 0 | |
| gpt-5.6-sol | high | failed | 7/8 | 31.1% | — | 0 | |
| gpt-5.6-luna | low | failed | 7/8 | 28.5% | — | 0 | |
| gpt-5.6-sol | none | failed | 7/8 | 25.2% | — | 0 | |
| claude-sonnet-5 | xhigh | failed | 2/8 | 3.4% | — | 0 |