1205generated behavioral tests
87.4%official best score
74.1%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 | 74.1% | ok | 893/1205 | $4.67 | 79 | 0.24h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 23.8% | ok | 287/1205 | n/a | 124 | 0.37h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 19.9% | ok | 240/1205 | n/a | 155 | 0.23h | 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
74.1% from 893/1205 ProgramBench-scored tests. Raw public eval statuses: error: 1, failure: 382, passed: 1051, skipped: 4.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
84620c2d0270eval.tests.test_additional_symbols.test_symbol_return (failure)84620c2d0270eval.tests.test_any_quantifier.test_lazy_any_of (failure)84620c2d0270eval.tests.test_boundary_and_special.test_vertical_tab (failure)84620c2d0270eval.tests.test_advanced_features.test_lazy_quantifier (failure)84620c2d0270eval.tests.test_compiler_errors.test_unrecognized_symbol_namespace (failure)84620c2d0270eval.tests.test_advanced_features.test_variables_basic (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 | Claude Opus 4.7 (xhigh) | Anthropic | 87.4% | $17.22 | 215 |
| 2 | GPT 5.5 (xhigh) | OpenAI | 84.9% | $4.22 | 56 |
| 3 | Claude Opus 4.7 | Anthropic | 78.9% | $5.37 | 135 |
| 4 | Claude Opus 4.6 | Anthropic | 74.4% | $12.05 | 325 |
| 5 | GPT 5.5 (high) | OpenAI | 71.5% | $3.79 | 54 |
| 6 | GPT 5.5 | OpenAI | 62.3% | $1.25 | 21 |
| 7 | Gemini 3 Flash | 62.3% | $0.28 | 138 | |
| 8 | GPT 5.4 | OpenAI | 34.2% | $0.45 | 19 |
| 9 | Claude Haiku 4.5 | Anthropic | 31.4% | $0.98 | 131 |
| 10 | GPT 5.4 mini | OpenAI | 6.4% | $0.03 | 9 |
| 11 | Gemini 3.1 Pro | 6.1% | $2.10 | 203 | |
| 12 | Claude Sonnet 4.6 | Anthropic | 3.4% | $16.67 | 502 |
| 13 | GPT 5 mini | OpenAI | 3.4% | $0.04 | 25 |