3034generated behavioral tests
31.0%official best score
14.5%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) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 14.5% | ok | 440/3034 | n/a | 90 | 0.35h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 0.0% | ok | 0/3034 | $6.76 | 118 | 0.27h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 0.0% | ok | 0/3034 | n/a | 76 | 0.22h | 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) · 20260519T215046Z
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260517T094700Z
0.0% from 0/3034 ProgramBench-scored tests. Raw public eval statuses: failure: 3038, skipped: 6.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch, interactive terminal behavior mismatch.
e1d7a4e20e53tests.test_advanced_features.test_proxy_preventextensions_trap (failure)e1d7a4e20e53tests.test_conversion_coercion_final_gaps.test_tonumber_string_formats (failure)e1d7a4e20e53tests.test_examples.test_circular_import (failure)e1d7a4e20e53tests.test_final_easy_wins.test_string_charat_null_this (failure)e1d7a4e20e53tests.test_error_path_sweep.test_date_invalid_string (failure)e1d7a4e20e53tests.test_import_attributes.test_basic_json_import_with_type_attribute (failure)
GPT 5.5 (xhigh) · 20260518T232006Z
Public eval evidence has not been exported for this row yet.
Official ProgramBench Results by Model
| # | Model | Provider | Score | Cost | Calls |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.7 (xhigh) | Anthropic | 31.0% | $31.65 | 261 |
| 2 | GPT 5.5 (xhigh) | OpenAI | 21.3% | $7.83 | 76 |
| 3 | GPT 5.5 (high) | OpenAI | 10.9% | $3.51 | 30 |
| 4 | GPT 5.5 | OpenAI | 4.0% | $1.67 | 26 |
| 5 | Claude Opus 4.6 | Anthropic | 3.6% | $16.39 | 248 |
| 6 | GPT 5.4 | OpenAI | 1.2% | $0.21 | 13 |
| 7 | Claude Opus 4.7 | Anthropic | 1.2% | $0.52 | 18 |
| 8 | Gemini 3.1 Pro | 1.0% | $1.30 | 119 | |
| 9 | Claude Sonnet 4.6 | Anthropic | 0.0% | $82.98 | 995 |
| 10 | Claude Haiku 4.5 | Anthropic | 0.0% | $0.55 | 124 |
| 11 | Gemini 3 Flash | 0.0% | $0.44 | 96 | |
| 12 | GPT 5.4 mini | OpenAI | 0.0% | $0.01 | 8 |
| 13 | GPT 5 mini | OpenAI | 0.0% | $0.01 | 10 |