13514generated behavioral tests
71.6%official best score
1.3%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 | 1.3% | ok | 182/13514 | $8.14 | 140 | 0.36h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 0.5% | ok | 62/13514 | n/a | 75 | 0.22h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 0.0% | ok | 5/13514 | n/a | 97 | 0.32h | 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
1.3% from 182/13514 ProgramBench-scored tests. Raw public eval statuses: failure: 85, not_run: 16525, passed: 191.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
281d3b6a53d3eval.tests.test_sqlite3.test_help_output (failure)281d3b6a53d3eval.tests.test_sqlite3.test_dot_dbinfo (failure)281d3b6a53d3eval.tests.test_sqlite3.test_dot_explain_on (failure)281d3b6a53d3eval.tests.test_sqlite3.test_dot_intck (failure)281d3b6a53d3eval.tests.test_sqlite3.test_help_short (failure)281d3b6a53d3eval.tests.test_sqlite3.test_dot_vfsinfo_with_db (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 | 71.6% | $14.62 | 155 |
| 2 | GPT 5.5 (high) | OpenAI | 68.4% | $3.06 | 45 |
| 3 | Claude Opus 4.7 | Anthropic | 67.0% | $1.08 | 37 |
| 4 | GPT 5.5 | OpenAI | 63.7% | $1.69 | 30 |
| 5 | Claude Haiku 4.5 | Anthropic | 20.3% | $0.57 | 137 |
| 6 | GPT 5 mini | OpenAI | 19.9% | $0.01 | 14 |
| 7 | Claude Sonnet 4.6 | Anthropic | 1.3% | $28.86 | 573 |
| 8 | Gemini 3 Flash | 0.9% | $0.19 | 74 | |
| 9 | Claude Opus 4.7 (xhigh) | Anthropic | 0.7% | $2.48 | 68 |
| 10 | GPT 5.4 | OpenAI | 0.6% | $0.20 | 9 |
| 11 | GPT 5.4 mini | OpenAI | 0.5% | $0.02 | 11 |
| 12 | Gemini 3.1 Pro | 0.4% | $0.85 | 77 | |
| 13 | Claude Opus 4.6 | Anthropic | 0.0% | $4.54 | 274 |