568generated behavioral tests
93.7%official best score
75.0%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) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 75.0% | ok | 426/568 | n/a | 64 | 0.24h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 60.9% | ok | 346/568 | n/a | 93 | 0.30h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 0.0% | error | 0/568 | $6.07 | 96 | 0.31h | manifest · eval json · eval summary · usage audit |
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) · 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.
GPT 5.5 (xhigh) · 20260517T094700Z
0.0% from 0/568 ProgramBench-scored tests. Raw public eval statuses: not_run: 733.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch.
800f83026235eval.tests.test_additional_coverage.test_all_styles_with_ascii (not_run)800f83026235eval.tests.test_additional_coverage.test_color_always_with_different_styles (not_run)800f83026235eval.tests.test_additional_coverage.test_debug_with_svg (not_run)800f83026235eval.tests.test_additional_coverage.test_format_combinations (not_run)800f83026235eval.tests.test_additional_coverage.test_format_conditional_minus (not_run)800f83026235eval.tests.test_additional_coverage.test_format_conditional_plus (not_run)
Official ProgramBench Results by Model
| # | Model | Provider | Score | Cost | Calls |
|---|---|---|---|---|---|
| 1 | GPT 5.5 (xhigh) | OpenAI | 93.7% | $12.72 | 109 |
| 2 | GPT 5.5 (high) | OpenAI | 88.0% | $3.35 | 42 |
| 3 | Claude Sonnet 4.6 | Anthropic | 79.6% | $49.86 | 658 |
| 4 | Claude Opus 4.7 (xhigh) | Anthropic | 78.9% | $4.01 | 117 |
| 5 | GPT 5.5 | OpenAI | 74.8% | $1.29 | 21 |
| 6 | Claude Haiku 4.5 | Anthropic | 58.8% | $1.17 | 166 |
| 7 | GPT 5.4 | OpenAI | 56.2% | $0.20 | 10 |
| 8 | Gemini 3.1 Pro | 35.6% | $0.94 | 49 | |
| 9 | GPT 5.4 mini | OpenAI | 27.6% | $0.07 | 20 |
| 10 | Claude Opus 4.7 | Anthropic | 3.3% | $7.75 | 160 |
| 11 | GPT 5 mini | OpenAI | 3.3% | $0.02 | 14 |
| 12 | Gemini 3 Flash | 1.2% | $0.24 | 49 | |
| 13 | Claude Opus 4.6 | Anthropic | 0.0% | $1.52 | 216 |