1166generated behavioral tests
85.1%official best score
78.9%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 | 78.9% | ok | 920/1166 | $4.28 | 60 | 0.22h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 74.7% | ok | 871/1166 | n/a | 48 | 0.20h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 54.2% | ok | 632/1166 | n/a | 57 | 0.23h | 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
78.9% from 920/1166 ProgramBench-scored tests. Raw public eval statuses: failure: 248, passed: 963, skipped: 3.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
48292bcdfab1eval.tests.test_edge_cases.test_invalid_text_color_value (failure)48292bcdfab1eval.tests.test_generation.test_generate_zsh_completion (failure)48292bcdfab1eval.tests.test_image_options.test_color_resolution_requires_image (failure)e809fcc6ed9btests.test_ascii.test_ascii_colors_invalid_out_of_bounds (failure)e809fcc6ed9btests.test_edge_cases.test_invalid_color_value (failure)e809fcc6ed9btests.test_image.test_color_resolution_without_image (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 | 85.1% | $12.76 | 109 |
| 2 | GPT 5.5 (high) | OpenAI | 81.9% | $4.06 | 45 |
| 3 | Claude Opus 4.6 | Anthropic | 81.7% | $16.25 | 289 |
| 4 | GPT 5.5 | OpenAI | 69.0% | $1.42 | 16 |
| 5 | Claude Opus 4.7 (xhigh) | Anthropic | 66.0% | $3.68 | 108 |
| 6 | Claude Opus 4.7 | Anthropic | 53.6% | $1.84 | 72 |
| 7 | GPT 5.4 | OpenAI | 52.2% | $0.22 | 8 |
| 8 | Claude Haiku 4.5 | Anthropic | 35.8% | $0.72 | 98 |
| 9 | GPT 5.4 mini | OpenAI | 15.8% | $0.03 | 7 |
| 10 | Gemini 3 Flash | 13.6% | $0.61 | 91 | |
| 11 | Claude Sonnet 4.6 | Anthropic | 2.3% | $33.45 | 604 |
| 12 | Gemini 3.1 Pro | 0.0% | $1.68 | 89 | |
| 13 | GPT 5 mini | OpenAI | 0.0% | $0.02 | 13 |