1852generated behavioral tests
80.1%official best score
82.3%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 | 82.3% | ok | 1525/1852 | $5.89 | 84 | 0.33h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 76.6% | ok | 1419/1852 | n/a | 71 | 0.20h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 21.5% | ok | 399/1852 | n/a | 81 | 0.25h | 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
82.3% from 1525/1852 ProgramBench-scored tests. Raw public eval statuses: failure: 335, passed: 1740, skipped: 1.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch.
bc9e04bf4f1dtests.test_basic.test_completions_zsh (failure)bc9e04bf4f1dtests.test_other_langs.test_typescript_variable_declarations (failure)5f48c669398ctests.test_basic_invocation.test_completions_zsh (failure)5f48c669398ctests.test_delete_squeeze.test_squeeze_newlines (failure)5f48c669398ctests.test_more_features.test_go_const (failure)5f48c669398ctests.test_more_features.test_typescript_type_params (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 | 80.1% | $10.26 | 91 |
| 2 | GPT 5.5 (high) | OpenAI | 79.8% | $3.89 | 40 |
| 3 | Claude Sonnet 4.6 | Anthropic | 69.5% | $19.30 | 565 |
| 4 | GPT 5.5 | OpenAI | 68.7% | $1.63 | 22 |
| 5 | Claude Opus 4.7 (xhigh) | Anthropic | 65.9% | $6.38 | 118 |
| 6 | Claude Opus 4.6 | Anthropic | 62.8% | $13.43 | 344 |
| 7 | GPT 5.4 | OpenAI | 58.1% | $0.41 | 12 |
| 8 | Gemini 3.1 Pro | 57.9% | $1.13 | 94 | |
| 9 | Gemini 3 Flash | 53.1% | $0.25 | 111 | |
| 10 | GPT 5 mini | OpenAI | 46.0% | $0.04 | 13 |
| 11 | Claude Haiku 4.5 | Anthropic | 43.6% | $1.15 | 167 |
| 12 | Claude Opus 4.7 | Anthropic | 2.4% | $1.57 | 64 |
| 13 | GPT 5.4 mini | OpenAI | 0.6% | $0.03 | 6 |