732generated behavioral tests
69.5%official best score
70.4%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) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 70.4% | ok | 515/732 | n/a | 41 | 0.17h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 65.8% | ok | 482/732 | $3.29 | 53 | 0.24h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 56.6% | ok | 414/732 | n/a | 55 | 0.19h | 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) · 20260518T232006Z
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260517T094700Z
65.8% from 482/732 ProgramBench-scored tests. Raw public eval statuses: error: 1, failure: 251, passed: 505, skipped: 3.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
62a430dc1ca2eval.tests.test_error_cases.test_negative_column_width (failure)62a430dc1ca2eval.tests.test_error_cases.test_invalid_column_width (failure)62a430dc1ca2eval.tests.test_input_output.test_input_short_flag (failure)62a430dc1ca2eval.tests.test_input_output.test_input_from_stdin (failure)62a430dc1ca2eval.tests.test_input_output.test_input_from_json_file (failure)d6cf5754a29ctests.test_cli_utils.test_embedded_languages_with_narrow_columns (failure)
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 | Claude Sonnet 4.6 | Anthropic | 69.5% | $21.30 | 409 |
| 2 | Claude Opus 4.7 (xhigh) | Anthropic | 62.6% | $11.82 | 171 |
| 3 | GPT 5.5 (high) | OpenAI | 60.9% | $2.96 | 27 |
| 4 | Claude Opus 4.7 | Anthropic | 54.8% | $3.25 | 95 |
| 5 | Gemini 3.1 Pro | 53.4% | $1.81 | 112 | |
| 6 | GPT 5.5 | OpenAI | 51.0% | $1.57 | 19 |
| 7 | Gemini 3 Flash | 40.6% | $0.26 | 71 | |
| 8 | GPT 5.4 | OpenAI | 32.0% | $0.39 | 11 |
| 9 | Claude Haiku 4.5 | Anthropic | 27.0% | $0.71 | 94 |
| 10 | GPT 5 mini | OpenAI | 12.2% | $0.01 | 8 |
| 11 | GPT 5.4 mini | OpenAI | 8.1% | $0.04 | 8 |
| 12 | GPT 5.5 (xhigh) | OpenAI | 2.0% | $10.01 | 98 |
| 13 | Claude Opus 4.6 | Anthropic | 0.0% | $2.82 | 193 |