755generated behavioral tests
90.2%official best score
88.2%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) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 88.2% | ok | 666/755 | $4.10 | 77 | 0.17h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 67.4% | ok | 509/755 | n/a | 91 | 0.22h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 47.7% | ok | 360/755 | n/a | 88 | 0.14h | 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
88.2% from 666/755 ProgramBench-scored tests. Raw public eval statuses: failure: 88, passed: 801, skipped: 5.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, terminal rendering mismatch.
0f7af0de4f7etests.test_cli.test_json_parse_error_empty_file (failure)0f7af0de4f7etests.test_cli.test_json_parse_error_missing_colon (failure)0f7af0de4f7etests.test_errors.test_truncated_json_unclosed_string (failure)0f7af0de4f7etests.test_cli.test_json_parse_error_unclosed_brace (failure)0f7af0de4f7etests.test_cli.test_json_parse_error_only_whitespace (failure)0f7af0de4f7etests.test_cli.test_json_parse_error_missing_array_closing_bracket (failure)
GPT 5.5 (xhigh) · 20260519T215046Z
Public eval evidence has not been exported for this row yet.
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 | 90.2% | $5.22 | 43 |
| 2 | GPT 5.5 (high) | OpenAI | 89.5% | $3.69 | 34 |
| 3 | Claude Opus 4.6 | Anthropic | 71.0% | $13.34 | 298 |
| 4 | GPT 5.5 | OpenAI | 68.6% | $0.96 | 14 |
| 5 | Claude Haiku 4.5 | Anthropic | 52.6% | $0.52 | 81 |
| 6 | Claude Sonnet 4.6 | Anthropic | 52.3% | $27.61 | 444 |
| 7 | Gemini 3.1 Pro | 46.8% | $1.11 | 60 | |
| 8 | Claude Opus 4.7 | Anthropic | 45.0% | $1.07 | 43 |
| 9 | GPT 5.4 mini | OpenAI | 44.4% | $0.03 | 7 |
| 10 | GPT 5.4 | OpenAI | 39.6% | $0.10 | 7 |
| 11 | GPT 5 mini | OpenAI | 35.6% | $0.01 | 10 |
| 12 | Gemini 3 Flash | 19.5% | $0.20 | 59 | |
| 13 | Claude Opus 4.7 (xhigh) | Anthropic | 9.4% | $4.62 | 116 |