304generated behavioral tests
92.1%official best score
90.5%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 | 90.5% | ok | 275/304 | $4.29 | 56 | 0.24h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 88.8% | ok | 270/304 | n/a | 71 | 0.20h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 56.2% | ok | 171/304 | n/a | 60 | 0.21h | 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
90.5% from 275/304 ProgramBench-scored tests. Raw public eval statuses: failure: 38, passed: 401, skipped: 1.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
a8a7a108c243eval.tests.test_basic_invocation.test_print_completions_powershell (failure)a8a7a108c243eval.tests.test_basic_invocation.test_print_completions_elvish (failure)a8a7a108c243eval.tests.test_file_operations.test_delete_directory (failure)a8a7a108c243eval.tests.test_ui_and_theming.test_qrcode_enabled (failure)ba5cd9ea5abeeval.tests.test_basic_invocation.test_print_completions_powershell (failure)ba5cd9ea5abeeval.tests.test_additional_coverage.test_multiple_interfaces (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 | 92.1% | $8.11 | 57 |
| 2 | GPT 5.5 (high) | OpenAI | 88.2% | $3.94 | 31 |
| 3 | GPT 5.5 | OpenAI | 80.3% | $1.24 | 16 |
| 4 | Claude Opus 4.6 | Anthropic | 78.6% | $9.68 | 137 |
| 5 | Claude Opus 4.7 (xhigh) | Anthropic | 77.3% | $2.94 | 68 |
| 6 | Claude Sonnet 4.6 | Anthropic | 76.6% | $25.03 | 420 |
| 7 | Claude Opus 4.7 | Anthropic | 61.8% | $0.94 | 30 |
| 8 | Gemini 3.1 Pro | 59.5% | $0.76 | 41 | |
| 9 | GPT 5.4 | OpenAI | 58.6% | $0.23 | 8 |
| 10 | Claude Haiku 4.5 | Anthropic | 45.4% | $0.62 | 107 |
| 11 | Gemini 3 Flash | 39.8% | $0.30 | 63 | |
| 12 | GPT 5.4 mini | OpenAI | 14.8% | $0.05 | 26 |
| 13 | GPT 5 mini | OpenAI | 9.5% | $0.03 | 13 |