1569generated behavioral tests
58.9%official best score
59.4%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) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 59.4% | ok | 932/1569 | $4.96 | 92 | 0.23h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 48.7% | ok | 764/1569 | n/a | 68 | 0.21h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 45.1% | ok | 708/1569 | n/a | 54 | 0.18h | 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
59.4% from 932/1569 ProgramBench-scored tests. Raw public eval statuses: failure: 798, passed: 1156, skipped: 4.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
ba586c2b191ceval.tests.test_query_advanced.test_query_flatten_function (failure)ba586c2b191ceval.tests.test_query_language.test_query_invalid_syntax (failure)1040f6b8a219tests.test_cli_commands_detail.test_completion_filters_by_extension (failure)1040f6b8a219tests.test_cli_query_advanced.test_builtin_sort (failure)1040f6b8a219tests.test_cli_setup_config.test_setup_completions_fish_with_existing_dir (failure)1040f6b8a219tests.test_cli_query_advanced.test_array_construction_multiple_queries (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 | GPT 5.5 (xhigh) | OpenAI | 58.9% | $7.32 | 69 |
| 2 | Claude Opus 4.6 | Anthropic | 55.1% | $20.28 | 431 |
| 3 | Claude Sonnet 4.6 | Anthropic | 53.3% | $29.34 | 585 |
| 4 | GPT 5.5 (high) | OpenAI | 49.1% | $3.31 | 33 |
| 5 | Claude Opus 4.7 | Anthropic | 45.4% | $2.54 | 66 |
| 6 | GPT 5.5 | OpenAI | 45.3% | $1.62 | 20 |
| 7 | Gemini 3.1 Pro | 35.6% | $2.30 | 130 | |
| 8 | Claude Haiku 4.5 | Anthropic | 32.4% | $0.86 | 144 |
| 9 | GPT 5.4 | OpenAI | 31.9% | $0.21 | 7 |
| 10 | Gemini 3 Flash | 27.5% | $0.31 | 117 | |
| 11 | GPT 5 mini | OpenAI | 21.7% | $0.07 | 34 |
| 12 | GPT 5.4 mini | OpenAI | 14.6% | $0.04 | 14 |
| 13 | Claude Opus 4.7 (xhigh) | Anthropic | 1.5% | $5.66 | 147 |