1095generated behavioral tests
88.2%official best score
89.6%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 | 89.6% | ok | 981/1095 | $4.99 | 80 | 0.26h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 49.4% | ok | 541/1095 | n/a | 69 | 0.21h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 49.0% | ok | 537/1095 | n/a | 64 | 0.20h | 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
89.6% from 981/1095 ProgramBench-scored tests. Raw public eval statuses: failure: 142, passed: 1150.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
0574468f33d5eval.tests.test_contexts_no_quoting.test_case_pattern_no_quoting (failure)0574468f33d5eval.tests.test_quoting_transformations.test_pwd_backtick_to_dollar_pwd (failure)0574468f33d5eval.tests.test_quoting_transformations.test_pwd_command_sub_not_in_assignment (failure)0574468f33d5eval.tests.test_test_transformations.test_xyes_pattern_removal_not_equals (failure)0574468f33d5eval.tests.test_edge_cases.test_process_substitution (failure)0574468f33d5eval.tests.test_error_handling.test_unexpected_eof_double_bracket (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 | 88.2% | $11.50 | 93 |
| 2 | GPT 5.5 (high) | OpenAI | 87.1% | $4.26 | 42 |
| 3 | Claude Opus 4.7 (xhigh) | Anthropic | 82.8% | $11.64 | 179 |
| 4 | Claude Sonnet 4.6 | Anthropic | 81.7% | $74.50 | 840 |
| 5 | GPT 5.5 | OpenAI | 79.7% | $1.16 | 18 |
| 6 | Claude Opus 4.7 | Anthropic | 72.0% | $1.99 | 69 |
| 7 | Gemini 3.1 Pro | 71.1% | $1.40 | 143 | |
| 8 | GPT 5.4 | OpenAI | 63.3% | $0.16 | 9 |
| 9 | Gemini 3 Flash | 59.6% | $0.24 | 90 | |
| 10 | Claude Haiku 4.5 | Anthropic | 54.4% | $0.96 | 147 |
| 11 | GPT 5.4 mini | OpenAI | 16.2% | $0.04 | 18 |
| 12 | GPT 5 mini | OpenAI | 7.2% | $0.03 | 14 |
| 13 | Claude Opus 4.6 | Anthropic | 0.0% | $2.13 | 256 |