Deepseek / deepseek-v4-flash
xhigh
91.2%
186 / 204 fixtures · 1 run(s)
48,618 input / 140,284 total output / 138,444 reasoning within output tokens $0.03948569
Reliability by Benchmark (Text)
Loading reliability summary…
Text vs JSON Schema Comparison
Pass Rate Delta
-17.2%
Text: 91.2% →
JSON: 74%
+8
Gained
JSON pass / text fail
−43
Lost
Text pass / JSON fail
143
Unchanged Pass
Both pass
10
Unchanged Fail
Both fail
Fixture Reliability Delta
| Fixture | Text | JSON | Delta |
|---|
Benchmark Deltas
| Benchmark | Text | JSON | Delta |
|---|---|---|---|
| commit_squash | 83.3% | 41.7% | -41.7% |
| git_grep | 100% | 66.7% | -33.3% |
| cherry_pick | 75% | 41.7% | -33.3% |
| blame_forensics | 83.3% | 58.3% | -25% |
| branch_cleanup | 100% | 75% | -25% |
| commit_messages | 100% | 75% | -25% |
| git_show | 100% | 75% | -25% |
| worktree_usage | 75% | 50% | -25% |
| git_bisect | 100% | 83.3% | -16.7% |
| git_clean | 100% | 83.3% | -16.7% |
| tag_management | 100% | 83.3% | -16.7% |
| submodule_usage | 83.3% | 91.7% | + 8.3% |
| git_log_format | 91.7% | 100% | + 8.3% |
| merge_conflicts | 83.3% | 75% | -8.3% |
| rebase | 75% | 66.7% | -8.3% |
| reflog | 100% | 91.7% | -8.3% |
| stash_recovery | 100% | 100% | + 0% |
Changed Fixtures (51)
f003 -loss f010 -loss f011 -loss f005 -loss f009 -loss f011 -loss f001 -loss f002 -loss f003 -loss f005 -loss f008 -loss f011 +gain f003 -loss f005 -loss f009 -loss f003 -loss f005 -loss f007 -loss f008 -loss f009 +gain f010 -loss f011 -loss f002 -loss f012 -loss f004 -loss f012 -loss f001 -loss f003 -loss f006 -loss f010 -loss f005 +gain f001 -loss f006 -loss f007 -loss f001 -loss f004 +gain f011 -loss f005 +gain f008 -loss f010 -loss f007 -loss f003 +gain f005 -loss f010 +gain f007 -loss f008 -loss f004 -loss f005 -loss f006 -loss f009 +gain f010 -loss
Fixture Gallery (204)
blame_forensics f001
0%
Failed
Tokens
302
Input
1,369
Total output
1,473
Reasoning within output
blame_forensics f002
100%
Passed
Tokens
273
Input
716
Total output
710
Reasoning within output
blame_forensics f003
100%
Passed
Tokens
305
Input
354
Total output
346
Reasoning within output
blame_forensics f004
100%
Passed
Tokens
243
Input
323
Total output
316
Reasoning within output
blame_forensics f005
0%
Failed
Tokens
301
Input
1,087
Total output
1,079
Reasoning within output
blame_forensics f006
100%
Passed
Tokens
288
Input
877
Total output
916
Reasoning within output
blame_forensics f007
100%
Passed
Tokens
268
Input
466
Total output
459
Reasoning within output
blame_forensics f008
100%
Passed
Tokens
288
Input
602
Total output
596
Reasoning within output
blame_forensics f009
100%
Passed
Tokens
171
Input
156
Total output
171
Reasoning within output
blame_forensics f010
100%
Passed
Tokens
445
Input
1,438
Total output
1,433
Reasoning within output
blame_forensics f011
100%
Passed
Tokens
283
Input
456
Total output
433
Reasoning within output
blame_forensics f012
100%
Passed
Tokens
206
Input
500
Total output
494
Reasoning within output
branch_cleanup f001
100%
Passed
Tokens
101
Input
204
Total output
187
Reasoning within output
branch_cleanup f002
100%
Passed
Tokens
190
Input
226
Total output
219
Reasoning within output
branch_cleanup f003
100%
Passed
Tokens
171
Input
517
Total output
585
Reasoning within output
branch_cleanup f004
100%
Passed
Tokens
199
Input
188
Total output
180
Reasoning within output
branch_cleanup f005
100%
Passed
Tokens
193
Input
288
Total output
283
Reasoning within output
branch_cleanup f006
100%
Passed
Tokens
195
Input
240
Total output
231
Reasoning within output
branch_cleanup f007
100%
Passed
Tokens
174
Input
274
Total output
322
Reasoning within output
branch_cleanup f008
100%
Passed
Tokens
198
Input
244
Total output
232
Reasoning within output
branch_cleanup f009
100%
Passed
Tokens
155
Input
176
Total output
197
Reasoning within output
branch_cleanup f010
100%
Passed
Tokens
116
Input
100
Total output
95
Reasoning within output
branch_cleanup f011
100%
Passed
Tokens
70
Input
56
Total output
64
Reasoning within output
branch_cleanup f012
100%
Passed
Tokens
227
Input
237
Total output
223
Reasoning within output
cherry_pick f001
100%
Passed
Tokens
171
Input
1,281
Total output
1,275
Reasoning within output
cherry_pick f002
100%
Passed
Tokens
199
Input
1,079
Total output
1,069
Reasoning within output
cherry_pick f003
100%
Passed
Tokens
89
Input
320
Total output
363
Reasoning within output
cherry_pick f004
0%
Failed
Tokens
226
Input
1,825
Total output
1,804
Reasoning within output
cherry_pick f005
100%
Passed
Tokens
228
Input
181
Total output
194
Reasoning within output
cherry_pick f006
100%
Passed
Tokens
215
Input
686
Total output
665
Reasoning within output
cherry_pick f007
100%
Passed
Tokens
208
Input
267
Total output
256
Reasoning within output
cherry_pick f008
100%
Passed
Tokens
202
Input
152
Total output
144
Reasoning within output
cherry_pick f009
100%
Passed
Tokens
187
Input
198
Total output
192
Reasoning within output
cherry_pick f010
100%
Passed
Tokens
290
Input
472
Total output
447
Reasoning within output
cherry_pick f011
0%
Failed
Tokens
240
Input
404
Total output
390
Reasoning within output
cherry_pick f012
0%
Failed
Tokens
343
Input
784
Total output
777
Reasoning within output
commit_messages f001
92.7%
Passed
Tokens
179
Input
1,134
Total output
1,041
Reasoning within output
commit_messages f002
84.3%
Passed
Tokens
275
Input
941
Total output
930
Reasoning within output
commit_messages f003
66%
Passed
Tokens
160
Input
433
Total output
373
Reasoning within output
commit_messages f004
91.7%
Passed
Tokens
187
Input
148
Total output
154
Reasoning within output
commit_messages f005
86.7%
Passed
Tokens
175
Input
529
Total output
519
Reasoning within output
commit_messages f006
85%
Passed
Tokens
130
Input
199
Total output
192
Reasoning within output
commit_messages f007
83%
Passed
Tokens
259
Input
340
Total output
319
Reasoning within output
commit_messages f008
91%
Passed
Tokens
145
Input
1,204
Total output
1,196
Reasoning within output
commit_messages f009
86.7%
Passed
Tokens
198
Input
934
Total output
925
Reasoning within output
commit_messages f010
92%
Passed
Tokens
345
Input
4,562
Total output
4,552
Reasoning within output
commit_messages f011
82%
Passed
Tokens
216
Input
1,550
Total output
1,541
Reasoning within output
commit_messages f012
75.7%
Passed
Tokens
199
Input
421
Total output
468
Reasoning within output
commit_squash f001
100%
Passed
Tokens
210
Input
830
Total output
816
Reasoning within output
commit_squash f002
100%
Passed
Tokens
194
Input
299
Total output
283
Reasoning within output
commit_squash f003
100%
Passed
Tokens
199
Input
602
Total output
588
Reasoning within output
commit_squash f004
0%
Failed
Tokens
192
Input
766
Total output
793
Reasoning within output
commit_squash f005
100%
Passed
Tokens
186
Input
953
Total output
944
Reasoning within output
commit_squash f006
100%
Passed
Tokens
172
Input
420
Total output
471
Reasoning within output
commit_squash f007
100%
Passed
Tokens
176
Input
177
Total output
163
Reasoning within output
commit_squash f008
100%
Passed
Tokens
187
Input
543
Total output
589
Reasoning within output
commit_squash f009
0%
Failed
Tokens
103
Input
0
Total output
315
Reasoning within output
commit_squash f010
100%
Passed
Tokens
190
Input
333
Total output
325
Reasoning within output
commit_squash f011
100%
Passed
Tokens
122
Input
215
Total output
198
Reasoning within output
commit_squash f012
100%
Passed
Tokens
200
Input
1,943
Total output
1,921
Reasoning within output
git_bisect f001
100%
Passed
Tokens
243
Input
125
Total output
119
Reasoning within output
git_bisect f002
100%
Passed
Tokens
278
Input
206
Total output
200
Reasoning within output
git_bisect f003
100%
Passed
Tokens
266
Input
292
Total output
287
Reasoning within output
git_bisect f004
100%
Passed
Tokens
282
Input
200
Total output
182
Reasoning within output
git_bisect f005
100%
Passed
Tokens
270
Input
177
Total output
172
Reasoning within output
git_bisect f006
100%
Passed
Tokens
297
Input
286
Total output
279
Reasoning within output
git_bisect f007
100%
Passed
Tokens
191
Input
94
Total output
75
Reasoning within output
git_bisect f008
100%
Passed
Tokens
262
Input
173
Total output
167
Reasoning within output
git_bisect f009
100%
Passed
Tokens
260
Input
169
Total output
163
Reasoning within output
git_bisect f010
100%
Passed
Tokens
187
Input
180
Total output
177
Reasoning within output
git_bisect f011
100%
Passed
Tokens
268
Input
217
Total output
186
Reasoning within output
git_bisect f012
100%
Passed
Tokens
293
Input
160
Total output
152
Reasoning within output
git_clean f001
100%
Passed
Tokens
128
Input
352
Total output
351
Reasoning within output
git_clean f002
100%
Passed
Tokens
138
Input
303
Total output
297
Reasoning within output
git_clean f003
100%
Passed
Tokens
136
Input
311
Total output
318
Reasoning within output
git_clean f004
100%
Passed
Tokens
136
Input
807
Total output
800
Reasoning within output
git_clean f005
100%
Passed
Tokens
126
Input
496
Total output
488
Reasoning within output
git_clean f006
100%
Passed
Tokens
153
Input
682
Total output
688
Reasoning within output
git_clean f007
100%
Passed
Tokens
148
Input
919
Total output
908
Reasoning within output
git_clean f008
100%
Passed
Tokens
151
Input
222
Total output
216
Reasoning within output
git_clean f009
100%
Passed
Tokens
146
Input
640
Total output
630
Reasoning within output
git_clean f010
100%
Passed
Tokens
142
Input
854
Total output
848
Reasoning within output
git_clean f011
100%
Passed
Tokens
139
Input
573
Total output
566
Reasoning within output
git_clean f012
100%
Passed
Tokens
142
Input
536
Total output
527
Reasoning within output
git_grep f001
100%
Passed
Tokens
131
Input
153
Total output
148
Reasoning within output
git_grep f002
100%
Passed
Tokens
142
Input
55
Total output
52
Reasoning within output
git_grep f003
100%
Passed
Tokens
163
Input
55
Total output
57
Reasoning within output
git_grep f004
100%
Passed
Tokens
72
Input
47
Total output
33
Reasoning within output
git_grep f005
100%
Passed
Tokens
186
Input
168
Total output
166
Reasoning within output
git_grep f006
100%
Passed
Tokens
37
Input
335
Total output
371
Reasoning within output
git_grep f007
100%
Passed
Tokens
156
Input
4,386
Total output
4,385
Reasoning within output
git_grep f008
100%
Passed
Tokens
136
Input
88
Total output
96
Reasoning within output
git_grep f009
100%
Passed
Tokens
129
Input
82
Total output
69
Reasoning within output
git_grep f010
100%
Passed
Tokens
139
Input
158
Total output
154
Reasoning within output
git_grep f011
100%
Passed
Tokens
193
Input
158
Total output
155
Reasoning within output
git_grep f012
100%
Passed
Tokens
143
Input
170
Total output
159
Reasoning within output
git_log_format f001
100%
Passed
Tokens
779
Input
125
Total output
122
Reasoning within output
git_log_format f002
100%
Passed
Tokens
821
Input
278
Total output
269
Reasoning within output
git_log_format f003
100%
Passed
Tokens
649
Input
148
Total output
137
Reasoning within output
git_log_format f004
100%
Passed
Tokens
786
Input
68
Total output
66
Reasoning within output
git_log_format f005
50%
Failed
Tokens
652
Input
510
Total output
556
Reasoning within output
git_log_format f006
100%
Passed
Tokens
782
Input
198
Total output
195
Reasoning within output
git_log_format f007
100%
Passed
Tokens
819
Input
118
Total output
112
Reasoning within output
git_log_format f008
100%
Passed
Tokens
1,014
Input
50
Total output
47
Reasoning within output
git_log_format f009
100%
Passed
Tokens
643
Input
85
Total output
83
Reasoning within output
git_log_format f010
100%
Passed
Tokens
1,133
Input
122
Total output
123
Reasoning within output
git_log_format f011
100%
Passed
Tokens
436
Input
180
Total output
176
Reasoning within output
git_log_format f012
100%
Passed
Tokens
412
Input
103
Total output
100
Reasoning within output
git_show f001
100%
Passed
Tokens
252
Input
98
Total output
91
Reasoning within output
git_show f002
100%
Passed
Tokens
242
Input
670
Total output
666
Reasoning within output
git_show f003
100%
Passed
Tokens
634
Input
235
Total output
227
Reasoning within output
git_show f004
100%
Passed
Tokens
467
Input
105
Total output
97
Reasoning within output
git_show f005
100%
Passed
Tokens
255
Input
136
Total output
132
Reasoning within output
git_show f006
100%
Passed
Tokens
261
Input
195
Total output
189
Reasoning within output
git_show f007
100%
Passed
Tokens
241
Input
110
Total output
104
Reasoning within output
git_show f008
100%
Passed
Tokens
257
Input
127
Total output
102
Reasoning within output
git_show f009
100%
Passed
Tokens
336
Input
247
Total output
217
Reasoning within output
git_show f010
100%
Passed
Tokens
248
Input
138
Total output
147
Reasoning within output
git_show f011
100%
Passed
Tokens
277
Input
104
Total output
109
Reasoning within output
git_show f012
100%
Passed
Tokens
371
Input
250
Total output
247
Reasoning within output
merge_conflicts f001
100%
Passed
Tokens
163
Input
736
Total output
846
Reasoning within output
merge_conflicts f002
100%
Passed
Tokens
188
Input
1,196
Total output
1,186
Reasoning within output
merge_conflicts f003
100%
Passed
Tokens
158
Input
5,355
Total output
5,346
Reasoning within output
merge_conflicts f004
0%
Failed
Tokens
213
Input
1,675
Total output
1,819
Reasoning within output
merge_conflicts f005
100%
Passed
Tokens
215
Input
394
Total output
376
Reasoning within output
merge_conflicts f006
100%
Passed
Tokens
202
Input
594
Total output
520
Reasoning within output
merge_conflicts f007
100%
Passed
Tokens
209
Input
395
Total output
384
Reasoning within output
merge_conflicts f008
100%
Passed
Tokens
185
Input
343
Total output
335
Reasoning within output
merge_conflicts f009
100%
Passed
Tokens
104
Input
214
Total output
209
Reasoning within output
merge_conflicts f010
100%
Passed
Tokens
257
Input
502
Total output
476
Reasoning within output
merge_conflicts f011
100%
Passed
Tokens
221
Input
345
Total output
322
Reasoning within output
merge_conflicts f012
0%
Failed
Tokens
310
Input
580
Total output
539
Reasoning within output
rebase f001
100%
Passed
Tokens
93
Input
422
Total output
418
Reasoning within output
rebase f002
100%
Passed
Tokens
206
Input
437
Total output
412
Reasoning within output
rebase f003
100%
Passed
Tokens
182
Input
237
Total output
258
Reasoning within output
rebase f004
0%
Failed
Tokens
224
Input
1,572
Total output
1,551
Reasoning within output
rebase f005
0%
Failed
Tokens
220
Input
913
Total output
1,062
Reasoning within output
rebase f006
100%
Passed
Tokens
214
Input
820
Total output
793
Reasoning within output
rebase f007
100%
Passed
Tokens
210
Input
255
Total output
245
Reasoning within output
rebase f008
100%
Passed
Tokens
198
Input
347
Total output
339
Reasoning within output
rebase f009
100%
Passed
Tokens
181
Input
87
Total output
79
Reasoning within output
rebase f010
100%
Passed
Tokens
282
Input
1,952
Total output
1,926
Reasoning within output
rebase f011
100%
Passed
Tokens
231
Input
696
Total output
670
Reasoning within output
rebase f012
0%
Failed
Tokens
266
Input
1,011
Total output
1,018
Reasoning within output
reflog f001
100%
Passed
Tokens
201
Input
339
Total output
316
Reasoning within output
reflog f002
100%
Passed
Tokens
278
Input
392
Total output
355
Reasoning within output
reflog f003
100%
Passed
Tokens
251
Input
929
Total output
849
Reasoning within output
reflog f004
100%
Passed
Tokens
359
Input
8,735
Total output
8,540
Reasoning within output
reflog f005
100%
Passed
Tokens
334
Input
1,807
Total output
1,666
Reasoning within output
reflog f006
100%
Passed
Tokens
296
Input
387
Total output
298
Reasoning within output
reflog f007
100%
Passed
Tokens
235
Input
115
Total output
59
Reasoning within output
reflog f008
100%
Passed
Tokens
395
Input
1,570
Total output
1,543
Reasoning within output
reflog f009
100%
Passed
Tokens
349
Input
9,147
Total output
9,048
Reasoning within output
reflog f010
100%
Passed
Tokens
241
Input
128
Total output
94
Reasoning within output
reflog f011
100%
Passed
Tokens
345
Input
2,747
Total output
2,644
Reasoning within output
reflog f012
100%
Passed
Tokens
351
Input
18,724
Total output
18,355
Reasoning within output
stash_recovery f001
100%
Passed
Tokens
203
Input
111
Total output
104
Reasoning within output
stash_recovery f002
100%
Passed
Tokens
268
Input
156
Total output
146
Reasoning within output
stash_recovery f003
100%
Passed
Tokens
111
Input
98
Total output
87
Reasoning within output
stash_recovery f004
100%
Passed
Tokens
189
Input
233
Total output
235
Reasoning within output
stash_recovery f005
100%
Passed
Tokens
119
Input
127
Total output
120
Reasoning within output
stash_recovery f006
100%
Passed
Tokens
246
Input
139
Total output
115
Reasoning within output
stash_recovery f007
100%
Passed
Tokens
195
Input
155
Total output
138
Reasoning within output
stash_recovery f008
100%
Passed
Tokens
242
Input
151
Total output
154
Reasoning within output
stash_recovery f009
100%
Passed
Tokens
277
Input
344
Total output
326
Reasoning within output
stash_recovery f010
100%
Passed
Tokens
197
Input
126
Total output
134
Reasoning within output
stash_recovery f011
100%
Passed
Tokens
187
Input
941
Total output
880
Reasoning within output
stash_recovery f012
100%
Passed
Tokens
206
Input
161
Total output
145
Reasoning within output
submodule_usage f001
100%
Passed
Tokens
40
Input
105
Total output
93
Reasoning within output
submodule_usage f002
100%
Passed
Tokens
170
Input
382
Total output
352
Reasoning within output
submodule_usage f003
75%
Failed
Tokens
172
Input
994
Total output
1,088
Reasoning within output
submodule_usage f004
100%
Passed
Tokens
78
Input
87
Total output
93
Reasoning within output
submodule_usage f005
100%
Passed
Tokens
126
Input
496
Total output
515
Reasoning within output
submodule_usage f006
100%
Passed
Tokens
157
Input
368
Total output
413
Reasoning within output
submodule_usage f007
100%
Passed
Tokens
127
Input
96
Total output
72
Reasoning within output
submodule_usage f008
100%
Passed
Tokens
165
Input
69
Total output
61
Reasoning within output
submodule_usage f009
100%
Passed
Tokens
159
Input
198
Total output
192
Reasoning within output
submodule_usage f010
0%
Failed
Tokens
171
Input
539
Total output
627
Reasoning within output
submodule_usage f011
100%
Passed
Tokens
175
Input
148
Total output
150
Reasoning within output
submodule_usage f012
100%
Passed
Tokens
125
Input
235
Total output
221
Reasoning within output
tag_management f001
100%
Passed
Tokens
129
Input
135
Total output
121
Reasoning within output
tag_management f002
100%
Passed
Tokens
130
Input
249
Total output
231
Reasoning within output
tag_management f003
100%
Passed
Tokens
136
Input
59
Total output
48
Reasoning within output
tag_management f004
100%
Passed
Tokens
140
Input
86
Total output
81
Reasoning within output
tag_management f005
100%
Passed
Tokens
152
Input
241
Total output
224
Reasoning within output
tag_management f006
100%
Passed
Tokens
135
Input
616
Total output
562
Reasoning within output
tag_management f007
100%
Passed
Tokens
139
Input
1,146
Total output
1,134
Reasoning within output
tag_management f008
100%
Passed
Tokens
66
Input
1,067
Total output
892
Reasoning within output
tag_management f009
100%
Passed
Tokens
144
Input
287
Total output
269
Reasoning within output
tag_management f010
100%
Passed
Tokens
45
Input
57
Total output
48
Reasoning within output
tag_management f011
100%
Passed
Tokens
45
Input
78
Total output
71
Reasoning within output
tag_management f012
100%
Passed
Tokens
154
Input
619
Total output
606
Reasoning within output
worktree_usage f001
100%
Passed
Tokens
201
Input
284
Total output
277
Reasoning within output
worktree_usage f002
100%
Passed
Tokens
209
Input
469
Total output
445
Reasoning within output
worktree_usage f003
100%
Passed
Tokens
263
Input
149
Total output
138
Reasoning within output
worktree_usage f004
100%
Passed
Tokens
264
Input
132
Total output
137
Reasoning within output
worktree_usage f005
100%
Passed
Tokens
196
Input
243
Total output
216
Reasoning within output
worktree_usage f006
100%
Passed
Tokens
206
Input
748
Total output
734
Reasoning within output
worktree_usage f007
100%
Passed
Tokens
199
Input
577
Total output
556
Reasoning within output
worktree_usage f008
0%
Failed
Tokens
222
Input
1,555
Total output
1,502
Reasoning within output
worktree_usage f009
0%
Failed
Tokens
298
Input
209
Total output
230
Reasoning within output
worktree_usage f010
100%
Passed
Tokens
273
Input
253
Total output
234
Reasoning within output
worktree_usage f011
100%
Passed
Tokens
269
Input
123
Total output
113
Reasoning within output
worktree_usage f012
0%
Failed
Tokens
196
Input
897
Total output
827
Reasoning within output