232generated behavioral tests
71.1%official best score
52.2%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 | 52.2% | ok | 121/232 | n/a | 55 | 0.18h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 48.7% | ok | 113/232 | $3.17 | 35 | 0.15h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 35.8% | ok | 83/232 | n/a | 41 | 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) · 20260519T215046Z
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260517T094700Z
48.7% from 113/232 ProgramBench-scored tests. Raw public eval statuses: failure: 121, passed: 116.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch.
bc90bea37ab6tests.test_cli.test_speed_factor_negative_rejected (failure)bc90bea37ab6tests.test_cli.test_help_short_flag (failure)bc90bea37ab6tests.test_error_paths.test_exit_after_modules_negative_rejected (failure)bc90bea37ab6tests.test_error_paths.test_instant_print_lines_negative_rejected (failure)bc90bea37ab6tests.test_core_coverage.test_mkinitcpio_os_release (failure)bc90bea37ab6tests.test_error_paths.test_modules_empty_string_rejected (failure)
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 (xhigh) | OpenAI | 71.1% | $5.22 | 42 |
| 2 | Claude Sonnet 4.6 | Anthropic | 59.1% | $17.93 | 298 |
| 3 | Claude Opus 4.6 | Anthropic | 53.0% | $11.11 | 158 |
| 4 | Claude Opus 4.7 (xhigh) | Anthropic | 48.7% | $3.83 | 71 |
| 5 | GPT 5.5 (high) | OpenAI | 36.6% | $2.91 | 24 |
| 6 | Gemini 3.1 Pro | 35.8% | $1.25 | 44 | |
| 7 | GPT 5.4 | OpenAI | 24.1% | $0.46 | 10 |
| 8 | Claude Haiku 4.5 | Anthropic | 24.1% | $0.58 | 79 |
| 9 | Claude Opus 4.7 | Anthropic | 23.3% | $1.70 | 49 |
| 10 | GPT 5.5 | OpenAI | 23.3% | $0.94 | 13 |
| 11 | Gemini 3 Flash | 20.7% | $0.47 | 48 | |
| 12 | GPT 5.4 mini | OpenAI | 12.5% | $0.02 | 10 |
| 13 | GPT 5 mini | OpenAI | 1.7% | $0.02 | 9 |