Openai / gpt-oss-20b

low

93.1% 190 / 204 fixtures · 1 run(s)
45,888 input / 18,508 total output / 12,189 reasoning within output tokens $0.00476041
79.9% 163 / 204 fixtures · 1 run(s)
46,588 input / 26,684 total output / 8,597 reasoning within output tokens $0.00557158
Loading reliability summary…
Pass Rate Delta
-13.2% Text: 93.1% → JSON: 79.9%
+4
Gained
JSON pass / text fail
−31
Lost
Text pass / JSON fail
159
Unchanged Pass
Both pass
10
Unchanged Fail
Both fail
Fixture Reliability Delta
Fixture Text JSON Delta
Benchmark Deltas
Benchmark Text JSON Delta
branch_cleanup 100% 50% -50%
git_log_format 100% 66.7% -33.3%
tag_management 100% 66.7% -33.3%
git_bisect 100% 75% -25%
git_show 100% 83.3% -16.7%
rebase 83.3% 66.7% -16.7%
submodule_usage 91.7% 83.3% -8.3%
cherry_pick 75% 66.7% -8.3%
commit_squash 100% 91.7% -8.3%
git_grep 100% 91.7% -8.3%
merge_conflicts 83.3% 75% -8.3%
worktree_usage 83.3% 75% -8.3%
blame_forensics 91.7% 91.7% + 0%
commit_messages 100% 100% + 0%
git_clean 83.3% 83.3% + 0%
reflog 91.7% 91.7% + 0%
stash_recovery 100% 100% + 0%
Changed Fixtures (35)