3050generated behavioral tests
7.0%official best score
9.9%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 | 9.9% | ok | 302/3050 | $5.34 | 83 | 0.22h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 8.8% | ok | 269/3050 | n/a | 54 | 0.19h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 8.2% | ok | 250/3050 | n/a | 78 | 0.26h | 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
9.9% from 302/3050 ProgramBench-scored tests. Raw public eval statuses: failure: 3161, passed: 398, skipped: 606.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch.
3f244e346e7etests.test_cmdutils_final.test_stream_specifier_program_id_no_value_error (failure)3f244e346e7etests.test_advanced_video.test_interlacing_flags (failure)3f244e346e7etests.test_ffmpeg_opt_depth.test_attach_file_with_metadata (failure)3f244e346e7etests.test_decoder_edges.test_decoder_ignore_err_flag (failure)3f244e346e7etests.test_encoder_remaining.test_interlaced_mpeg4_encoding (failure)3f244e346e7etests.test_filter_graph_edges.test_empty_filter_string_error (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 | 7.0% | $3.59 | 32 |
| 2 | GPT 5.5 | OpenAI | 6.1% | $1.16 | 15 |
| 3 | Gemini 3.1 Pro | 5.3% | $0.92 | 52 | |
| 4 | GPT 5.4 | OpenAI | 4.6% | $0.39 | 12 |
| 5 | Claude Sonnet 4.6 | Anthropic | 4.5% | $12.51 | 273 |
| 6 | GPT 5.5 (high) | OpenAI | 4.2% | $2.60 | 28 |
| 7 | GPT 5.4 mini | OpenAI | 3.9% | $0.03 | 13 |
| 8 | Gemini 3 Flash | 3.5% | $0.38 | 89 | |
| 9 | Claude Opus 4.7 | Anthropic | 3.3% | $1.07 | 35 |
| 10 | Claude Haiku 4.5 | Anthropic | 2.8% | $0.48 | 103 |
| 11 | Claude Opus 4.7 (xhigh) | Anthropic | 2.4% | $1.50 | 32 |
| 12 | GPT 5 mini | OpenAI | 2.2% | $0.02 | 15 |
| 13 | Claude Opus 4.6 | Anthropic | 0.0% | $2.06 | 221 |