291generated behavioral tests
77.0%official best score
60.8%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) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 60.8% | ok | 177/291 | n/a | 109 | 0.30h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 0.0% | ok | 0/291 | $4.75 | 76 | 0.25h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 0.0% | ok | 0/291 | n/a | 42 | 0.16h | 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) · 20260519T215046Z
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260517T094700Z
0.0% from 0/291 ProgramBench-scored tests. Raw public eval statuses: not_run: 298.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
bc90bea37ab6tests.test_command_timer.test_command_name_count_mismatch_error (not_run)bc90bea37ab6tests.test_command_timer.test_command_with_backslash_escape (not_run)bc90bea37ab6tests.test_command_timer.test_command_with_double_quotes_preserved (not_run)bc90bea37ab6tests.test_command_timer.test_command_with_multiple_pipes_and_redirects (not_run)bc90bea37ab6tests.test_command_timer.test_command_with_shell_metacharacters (not_run)bc90bea37ab6tests.test_command_timer.test_duplicate_parameter_names_error (not_run)
GPT 5.5 (xhigh) · 20260518T232006Z
Public eval evidence has not been exported for this row yet.
Official ProgramBench Results by Model
| # | Model | Provider | Score | Cost | Calls |
|---|---|---|---|---|---|
| 1 | GPT 5.5 (high) | OpenAI | 77.0% | $2.48 | 25 |
| 2 | Claude Opus 4.7 | Anthropic | 54.3% | $1.55 | 41 |
| 3 | GPT 5.5 | OpenAI | 49.1% | $1.37 | 18 |
| 4 | GPT 5.4 | OpenAI | 48.8% | $0.25 | 9 |
| 5 | Gemini 3.1 Pro | 43.6% | $1.29 | 55 | |
| 6 | Claude Haiku 4.5 | Anthropic | 24.1% | $0.58 | 85 |
| 7 | Claude Sonnet 4.6 | Anthropic | 19.2% | $31.21 | 546 |
| 8 | Gemini 3 Flash | 15.5% | $0.27 | 68 | |
| 9 | Claude Opus 4.6 | Anthropic | 11.7% | $13.58 | 265 |
| 10 | GPT 5.4 mini | OpenAI | 7.2% | $0.05 | 10 |
| 11 | GPT 5 mini | OpenAI | 0.7% | $0.02 | 9 |
| 12 | GPT 5.5 (xhigh) | OpenAI | 0.0% | $13.44 | 107 |
| 13 | Claude Opus 4.7 (xhigh) | Anthropic | n/a | $13.94 | 180 |