531generated behavioral tests
87.2%official best score
75.9%Codex best score
3Codex result rows
Codex Results by Model
| # | Model | Run | Agent | Mode | Score | Eval | Tests | Est. cost | Calls | Wall | Evidence |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 75.9% | ok | 403/531 | $3.69 | 61 | 0.22h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 71.4% | ok | 379/531 | n/a | 87 | 0.21h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 68.7% | ok | 365/531 | n/a | 74 | 0.21h | not exported |
Why Scores Differ
Official ProgramBench rows are the public mini-SWE-agent submissions. GoalBench rows are separate Codex /goal submissions. Same model label does not mean the same agent, prompt, tool loop, or generated implementation. A GoalBench row is resolved only when the ProgramBench-scored pass rate is exactly 100%.
GPT 5.5 (xhigh) · 20260517T094700Z
75.9% from 403/531 ProgramBench-scored tests. Raw public eval statuses: failure: 132, passed: 445.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch, interactive terminal behavior mismatch.
a9c8e0729570eval.tests.test_zoxide.TestImportCommand.test_import_from_fasd_alias (failure)a9c8e0729570eval.tests.test_zoxide.TestInitCommand.test_init_posix_alias_ksh (failure)a9c8e0729570eval.tests.test_zoxide.TestInitCommand.test_init_no_aliases_is_alias_for_no_cmd (failure)a9c8e0729570eval.tests.test_zoxide.TestEditCommand.test_edit_reload_subcommand (failure)a9c8e0729570eval.tests.test_zoxide.TestEditCommand.test_edit_delete_subcommand (failure)a9c8e0729570eval.tests.test_zoxide.TestEditCommand.test_edit_decrement_subcommand (failure)
GPT 5.5 (xhigh) · 20260518T232006Z
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260519T215046Z
Public eval evidence has not been exported for this row yet.
Official ProgramBench Results by Model
| # | Model | Provider | Score | Cost | Calls |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.7 (xhigh) | Anthropic | 87.2% | $32.86 | 378 |
| 2 | GPT 5.5 (xhigh) | OpenAI | 83.4% | $10.70 | 98 |
| 3 | GPT 5.5 (high) | OpenAI | 79.8% | $4.39 | 47 |
| 4 | Claude Opus 4.7 | Anthropic | 76.5% | $10.07 | 187 |
| 5 | Claude Sonnet 4.6 | Anthropic | 69.9% | $14.61 | 368 |
| 6 | GPT 5.5 | OpenAI | 65.0% | $0.95 | 16 |
| 7 | Claude Opus 4.6 | Anthropic | 63.7% | $18.43 | 307 |
| 8 | Gemini 3.1 Pro | 58.2% | $2.10 | 121 | |
| 9 | GPT 5.4 | OpenAI | 53.5% | $0.34 | 9 |
| 10 | Gemini 3 Flash | 52.0% | $0.42 | 157 | |
| 11 | Claude Haiku 4.5 | Anthropic | 46.7% | $0.88 | 112 |
| 12 | GPT 5.4 mini | OpenAI | 1.5% | $0.03 | 8 |
| 13 | GPT 5 mini | OpenAI | 0.0% | $0.05 | 15 |