657generated behavioral tests
25.4%official best score
26.5%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 | 26.5% | ok | 174/657 | n/a | 51 | 0.22h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 21.2% | ok | 139/657 | $5.66 | 67 | 0.27h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 14.0% | ok | 92/657 | n/a | 56 | 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
21.2% from 139/657 ProgramBench-scored tests. Raw public eval statuses: failure: 888, passed: 429, skipped: 2.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
14d8b1f8ec41eval.tests.test_dataset_advanced.test_dataset_check_vector (failure)14d8b1f8ec41eval.tests.test_pipeline_advanced.test_vector_pipeline_segmentize (failure)14d8b1f8ec41eval.tests.test_mdim.test_vector_sql_multiple (failure)14d8b1f8ec41eval.tests.test_basic.test_filename_shortcut_vector (failure)14d8b1f8ec41eval.tests.test_info_shortcut.test_convert_shortcut_vector (failure)14d8b1f8ec41eval.tests.test_more_coverage.test_raster_tile_with_title (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 | Claude Sonnet 4.6 | Anthropic | 25.4% | $48.10 | 684 |
| 2 | GPT 5.5 (high) | OpenAI | 17.2% | $3.46 | 36 |
| 3 | GPT 5.5 | OpenAI | 11.9% | $0.94 | 14 |
| 4 | Claude Opus 4.7 (xhigh) | Anthropic | 10.0% | $8.27 | 154 |
| 5 | Gemini 3 Flash | 6.7% | $0.24 | 70 | |
| 6 | Claude Haiku 4.5 | Anthropic | 6.1% | $0.64 | 142 |
| 7 | GPT 5.4 | OpenAI | 5.3% | $0.36 | 9 |
| 8 | Gemini 3.1 Pro | 5.0% | $0.74 | 39 | |
| 9 | GPT 5.4 mini | OpenAI | 5.0% | $0.03 | 7 |
| 10 | Claude Opus 4.7 | Anthropic | 3.7% | $0.47 | 22 |
| 11 | GPT 5 mini | OpenAI | 1.7% | $0.09 | 35 |
| 12 | GPT 5.5 (xhigh) | OpenAI | 0.0% | $4.39 | 41 |
| 13 | Claude Opus 4.6 | Anthropic | 0.0% | $2.60 | 218 |