522generated behavioral tests
84.5%official best score
63.2%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 | 63.2% | ok | 330/522 | n/a | 67 | 0.18h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 55.2% | ok | 288/522 | $2.64 | 54 | 0.16h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 51.1% | ok | 267/522 | n/a | 97 | 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) · 20260518T232006Z
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260517T094700Z
55.2% from 288/522 ProgramBench-scored tests. Raw public eval statuses: failure: 269, passed: 328, skipped: 2.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
66fdfa416565eval.tests.test_argument_parsing.test_client_upload_size_formats (failure)66fdfa416565eval.tests.test_argument_parsing.test_client_custom_host (failure)66fdfa416565eval.tests.test_argument_parsing.test_server_buffer_size_with_M (failure)66fdfa416565eval.tests.test_argument_parsing.test_server_buffer_size_plain_number (failure)66fdfa416565eval.tests.test_advanced_options.test_server_with_initial_rtt (failure)66fdfa416565eval.tests.test_advanced_options.test_server_with_multiple_options (failure)
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 | 84.5% | $3.83 | 40 |
| 2 | GPT 5.5 (high) | OpenAI | 73.0% | $3.09 | 34 |
| 3 | Claude Opus 4.7 (xhigh) | Anthropic | 63.4% | $5.00 | 117 |
| 4 | GPT 5.5 | OpenAI | 62.3% | $1.31 | 18 |
| 5 | Claude Opus 4.6 | Anthropic | 61.7% | $7.44 | 147 |
| 6 | Claude Sonnet 4.6 | Anthropic | 57.3% | $16.65 | 375 |
| 7 | Gemini 3.1 Pro | 56.5% | $1.99 | 88 | |
| 8 | Claude Opus 4.7 | Anthropic | 53.8% | $2.27 | 54 |
| 9 | GPT 5.4 | OpenAI | 46.9% | $0.31 | 10 |
| 10 | Gemini 3 Flash | 26.2% | $0.32 | 88 | |
| 11 | Claude Haiku 4.5 | Anthropic | 26.2% | $0.51 | 92 |
| 12 | GPT 5.4 mini | OpenAI | 20.9% | $0.04 | 9 |
| 13 | GPT 5 mini | OpenAI | 19.3% | $0.02 | 9 |