224generated behavioral tests
90.2%official best score
88.8%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 | 88.8% | ok | 199/224 | $4.05 | 84 | 0.21h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 76.3% | ok | 171/224 | n/a | 80 | 0.20h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 17.0% | ok | 38/224 | n/a | 61 | 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
88.8% from 199/224 ProgramBench-scored tests. Raw public eval statuses: failure: 26, passed: 207.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, interactive terminal behavior mismatch.
bc90bea37ab6tests.test_gaps.test_invalid_quoted_key_in_ungron (failure)bc90bea37ab6tests.test_gron_core.test_truncated_json_error (failure)bc90bea37ab6tests.test_gaps.test_invalid_json_statement_array_length (failure)bc90bea37ab6tests.test_gaps.test_colorize_json_output_ungron (failure)bc90bea37ab6tests.test_harvest.test_ungron_grep_separators_gron (failure)bc90bea37ab6tests.test_integration.test_ungron_negative_array_index_error (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 | Claude Opus 4.7 | Anthropic | 90.2% | $9.93 | 206 |
| 2 | Claude Opus 4.7 (xhigh) | Anthropic | 88.8% | $14.90 | 217 |
| 3 | Claude Opus 4.6 | Anthropic | 87.9% | $15.33 | 358 |
| 4 | GPT 5.5 (high) | OpenAI | 87.1% | $2.70 | 44 |
| 5 | GPT 5.5 (xhigh) | OpenAI | 85.7% | $4.82 | 46 |
| 6 | Claude Sonnet 4.6 | Anthropic | 84.4% | $14.17 | 459 |
| 7 | Gemini 3 Flash | 79.0% | $0.31 | 121 | |
| 8 | GPT 5.5 | OpenAI | 77.7% | $1.19 | 25 |
| 9 | Gemini 3.1 Pro | 71.0% | $1.56 | 177 | |
| 10 | Claude Haiku 4.5 | Anthropic | 70.5% | $0.69 | 107 |
| 11 | GPT 5.4 | OpenAI | 58.0% | $0.36 | 13 |
| 12 | GPT 5 mini | OpenAI | 22.3% | $0.03 | 11 |
| 13 | GPT 5.4 mini | OpenAI | 20.1% | $0.03 | 10 |