5228generated behavioral tests
14.7%official best score
17.0%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 | 17.0% | ok | 889/5228 | $6.38 | 91 | 0.36h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 15.7% | ok | 819/5228 | n/a | 41 | 0.21h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 14.9% | ok | 779/5228 | n/a | 66 | 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
17.0% from 889/5228 ProgramBench-scored tests. Raw public eval statuses: failure: 4524, passed: 942, skipped: 1.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
08afe9ff1778eval.tests.test_pandoc.test_bash_completion (failure)08afe9ff1778eval.tests.test_pandoc.test_markdown_to_org (failure)08afe9ff1778eval.tests.test_pandoc.test_markdown_to_mediawiki (failure)08afe9ff1778eval.tests.test_pandoc.test_print_default_data_file (failure)08afe9ff1778eval.tests.test_pandoc.test_markdown_to_dokuwiki (failure)08afe9ff1778eval.tests.test_pandoc.test_section_divs (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 (high) | OpenAI | 14.7% | $2.59 | 24 |
| 2 | Claude Sonnet 4.6 | Anthropic | 14.1% | $35.83 | 568 |
| 3 | Claude Opus 4.7 (xhigh) | Anthropic | 10.7% | $2.53 | 83 |
| 4 | GPT 5.5 | OpenAI | 10.1% | $1.63 | 21 |
| 5 | Gemini 3.1 Pro | 6.3% | $1.33 | 95 | |
| 6 | Gemini 3 Flash | 6.2% | $0.20 | 67 | |
| 7 | GPT 5.4 | OpenAI | 5.2% | $0.24 | 8 |
| 8 | Claude Opus 4.7 | Anthropic | 4.7% | $1.04 | 46 |
| 9 | GPT 5.4 mini | OpenAI | 0.6% | $0.05 | 19 |
| 10 | GPT 5 mini | OpenAI | 0.4% | $0.06 | 20 |
| 11 | Claude Opus 4.6 | Anthropic | 0.1% | $17.70 | 339 |
| 12 | Claude Haiku 4.5 | Anthropic | 0.1% | $0.90 | 105 |
| 13 | GPT 5.5 (xhigh) | OpenAI | 0.0% | $3.56 | 41 |