631generated behavioral tests
61.3%official best score
55.2%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 | 55.2% | ok | 348/631 | n/a | 89 | 0.29h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 47.1% | ok | 297/631 | $5.17 | 82 | 0.28h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 43.3% | ok | 273/631 | n/a | 83 | 0.29h | 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
47.1% from 297/631 ProgramBench-scored tests. Raw public eval statuses: failure: 336, passed: 375, skipped: 3.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch.
a06931aa0fbaeval.tests.test_jp2a.test_version_output (failure)a06931aa0fbaeval.tests.test_jp2a.test_help_output (failure)a06931aa0fbaeval.tests.test_jp2a.test_debug_mode (failure)a06931aa0fbaeval.tests.test_jp2a.test_custom_weights_sum_validation (failure)a06931aa0fbaeval.tests.test_jp2a_coverage.test_invalid_width_zero (failure)a06931aa0fbaeval.tests.test_jp2a_coverage.test_invalid_height_zero (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 | 61.3% | $8.75 | 97 |
| 2 | Claude Opus 4.7 | Anthropic | 56.1% | $16.11 | 239 |
| 3 | GPT 5.5 | OpenAI | 45.0% | $1.50 | 28 |
| 4 | GPT 5.5 (high) | OpenAI | 44.4% | $5.11 | 46 |
| 5 | Claude Sonnet 4.6 | Anthropic | 42.2% | $45.93 | 495 |
| 6 | Claude Opus 4.6 | Anthropic | 42.0% | $13.07 | 192 |
| 7 | Gemini 3 Flash | 33.3% | $0.31 | 68 | |
| 8 | Claude Haiku 4.5 | Anthropic | 30.0% | $1.24 | 162 |
| 9 | Gemini 3.1 Pro | 26.0% | $2.21 | 100 | |
| 10 | GPT 5.4 | OpenAI | 25.2% | $0.43 | 15 |
| 11 | Claude Opus 4.7 (xhigh) | Anthropic | 19.2% | $15.72 | 169 |
| 12 | GPT 5.4 mini | OpenAI | 14.1% | $0.05 | 37 |
| 13 | GPT 5 mini | OpenAI | 4.6% | $0.01 | 12 |