1438generated behavioral tests
82.0%official best score
72.7%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 | 72.7% | ok | 1045/1438 | $7.28 | 129 | 0.35h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 69.2% | ok | 995/1438 | n/a | 62 | 0.20h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 68.4% | ok | 983/1438 | n/a | 87 | 0.24h | 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
72.7% from 1045/1438 ProgramBench-scored tests. Raw public eval statuses: failure: 416, passed: 1487, skipped: 2.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch.
470d1be7dca2eval.tests.test_basic.test_warning_list (failure)470d1be7dca2eval.tests.test_basic.test_debug_list (failure)470d1be7dca2eval.tests.test_basic.test_help_flag (failure)470d1be7dca2eval.tests.test_error_handling.test_circular_dependency (failure)470d1be7dca2eval.tests.test_more_subtools.test_warning_phonycycle_err (failure)470d1be7dca2eval.tests.test_basic.test_help_short_flag (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 | 82.0% | $15.17 | 144 |
| 2 | Claude Opus 4.6 | Anthropic | 72.3% | $16.10 | 324 |
| 3 | Claude Opus 4.7 (xhigh) | Anthropic | 68.3% | $14.14 | 212 |
| 4 | Claude Sonnet 4.6 | Anthropic | 67.5% | $18.29 | 400 |
| 5 | GPT 5.5 (high) | OpenAI | 65.5% | $3.50 | 45 |
| 6 | GPT 5.5 | OpenAI | 57.2% | $0.90 | 15 |
| 7 | Gemini 3 Flash | 42.3% | $0.23 | 65 | |
| 8 | Gemini 3.1 Pro | 37.1% | $0.68 | 45 | |
| 9 | Claude Haiku 4.5 | Anthropic | 34.2% | $0.91 | 129 |
| 10 | GPT 5 mini | OpenAI | 5.6% | $0.01 | 11 |
| 11 | Claude Opus 4.7 | Anthropic | 4.8% | $1.69 | 69 |
| 12 | GPT 5.4 | OpenAI | 3.5% | $0.20 | 9 |
| 13 | GPT 5.4 mini | OpenAI | 0.0% | $0.03 | 14 |