668generated behavioral tests
68.4%official best score
65.0%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 | 65.0% | ok | 434/668 | $6.14 | 94 | 0.31h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 62.4% | ok | 417/668 | n/a | 66 | 0.20h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 57.8% | ok | 386/668 | n/a | 57 | 0.22h | 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
65.0% from 434/668 ProgramBench-scored tests. Raw public eval statuses: failure: 276, passed: 574, skipped: 3.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
768b1adb0770eval.tests.test_comprehensive_generation.test_complex_markdown_features (failure)768b1adb0770eval.tests.test_comprehensive_generation.test_full_blog_with_all_features (failure)94c38132ca06eval.tests.test_additional_features.test_custom_css_js_files (failure)94c38132ca06eval.tests.test_additional_features.test_post_with_special_characters_in_title (failure)94c38132ca06eval.tests.test_additional_features.test_long_post_title (failure)94c38132ca06eval.tests.test_content_handling.test_page_without_date (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 | GPT 5.5 (xhigh) | OpenAI | 68.4% | $12.42 | 97 |
| 2 | GPT 5.5 (high) | OpenAI | 53.7% | $2.80 | 29 |
| 3 | GPT 5.5 | OpenAI | 50.3% | $1.29 | 14 |
| 4 | Claude Opus 4.6 | Anthropic | 45.4% | $13.96 | 206 |
| 5 | GPT 5 mini | OpenAI | 27.8% | $0.02 | 13 |
| 6 | Claude Opus 4.7 (xhigh) | Anthropic | 24.6% | $1.26 | 42 |
| 7 | GPT 5.4 | OpenAI | 22.5% | $0.33 | 11 |
| 8 | Gemini 3.1 Pro | 22.3% | $1.26 | 62 | |
| 9 | Claude Opus 4.7 | Anthropic | 21.4% | $0.74 | 28 |
| 10 | Gemini 3 Flash | 18.6% | $0.35 | 79 | |
| 11 | Claude Sonnet 4.6 | Anthropic | 11.2% | $57.18 | 790 |
| 12 | GPT 5.4 mini | OpenAI | 8.2% | $0.06 | 14 |
| 13 | Claude Haiku 4.5 | Anthropic | 0.0% | $0.94 | 115 |