335generated behavioral tests
99.7%official best score
98.8%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 | 98.8% | ok | 331/335 | $4.74 | 86 | 0.27h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 71.3% | ok | 239/335 | n/a | 74 | 0.23h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 71.0% | ok | 238/335 | n/a | 82 | 0.43h | 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
98.8% from 331/335 ProgramBench-scored tests. Raw public eval statuses: failure: 4, passed: 343, skipped: 1.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
a7a2f5f083bfeval.tests.test_csview.test_line_numbers_seq_alias (failure)a7a2f5f083bfeval.tests.test_csview.test_style_case_insensitive (failure)d31025219260tests.test_input.test_emoji_wide_unicode_characters (failure)65af426d9fb3eval.tests.test_argparse.test_flag_value_forms_are_equivalent_for_success (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 | 99.7% | $4.82 | 62 |
| 2 | Claude Opus 4.7 (xhigh) | Anthropic | 98.2% | $11.47 | 167 |
| 3 | GPT 5.5 (high) | OpenAI | 97.3% | $2.55 | 35 |
| 4 | Claude Opus 4.6 | Anthropic | 96.1% | $8.20 | 249 |
| 5 | GPT 5.5 | OpenAI | 93.1% | $0.75 | 12 |
| 6 | Gemini 3.1 Pro | 93.1% | $1.37 | 136 | |
| 7 | Claude Opus 4.7 | Anthropic | 92.5% | $3.38 | 68 |
| 8 | Claude Sonnet 4.6 | Anthropic | 92.5% | $7.72 | 285 |
| 9 | GPT 5.4 | OpenAI | 88.7% | $0.16 | 8 |
| 10 | Gemini 3 Flash | 83.0% | $0.28 | 100 | |
| 11 | GPT 5.4 mini | OpenAI | 48.1% | $0.05 | 9 |
| 12 | Claude Haiku 4.5 | Anthropic | 0.6% | $0.63 | 109 |
| 13 | GPT 5 mini | OpenAI | 0.0% | $0.02 | 14 |