evalkit
Sample

Sample output produced by evalkit against local fixture endpoints. These numbers are not a benchmark of any real model. Latency is the localhost round trip from the run that wrote this file. Token counts are the usage field returned by the fixture, not a tokenizer. Cost multiplies those counts by the illustrative rates stored with the run.

Back to sample

json_extract

99eeb96a42e540ea8d9653ce1d8a30a4

fixture-exact

n 8 · exact 8/8 · schema 8/8

damagedexact 1.000 · schema 1.000 · judge 1.000

Customer Jane Doe (jane@example.com) says order #A-1042 arrived damaged.

21.8 ms · 65 in / 25 out · $2.47e-5

FieldExpectedPredictedMatch
nameJane DoeJane Doematch
emailjane@example.comjane@example.commatch
order_idA-1042A-1042match
issuearrived damagedarrived damagedmatch

Judge: field overlap 1.000

{
"name": "Jane Doe",
"email": "jane@example.com",
"order_id": "A-1042",
"issue": "arrived damaged"
}
missingexact 1.000 · schema 1.000 · judge 1.000

Luis Ortega (luis.ortega@example.com) cannot find order #B-2201; the item is missing.

11.5 ms · 68 in / 27 out · $2.64e-5

FieldExpectedPredictedMatch
nameLuis OrtegaLuis Ortegamatch
emailluis.ortega@example.comluis.ortega@example.commatch
order_idB-2201B-2201match
issuemissing itemmissing itemmatch

Judge: field overlap 1.000

{
"name": "Luis Ortega",
"email": "luis.ortega@example.com",
"order_id": "B-2201",
"issue": "missing item"
}
billingexact 1.000 · schema 1.000 · judge 1.000

Priya Shah (priya.shah@example.com) was charged twice for order #C-118.

9.4 ms · 65 in / 26 out · $2.53e-5

FieldExpectedPredictedMatch
namePriya ShahPriya Shahmatch
emailpriya.shah@example.compriya.shah@example.commatch
order_idC-118C-118match
issuecharged twicecharged twicematch

Judge: field overlap 1.000

{
"name": "Priya Shah",
"email": "priya.shah@example.com",
"order_id": "C-118",
"issue": "charged twice"
}
addressexact 1.000 · schema 1.000 · judge 1.000

Kenji Sato (kenji.sato@example.com) received order #D-9 at the wrong address.

14.1 ms · 66 in / 26 out · $2.55e-5

FieldExpectedPredictedMatch
nameKenji SatoKenji Satomatch
emailkenji.sato@example.comkenji.sato@example.commatch
order_idD-9D-9match
issuewrong addresswrong addressmatch

Judge: field overlap 1.000

{
"name": "Kenji Sato",
"email": "kenji.sato@example.com",
"order_id": "D-9",
"issue": "wrong address"
}
lateexact 1.000 · schema 1.000 · judge 1.000

Amira Hassan (amira.hassan@example.com) is still waiting on order #E-440; delivery is late.

12.1 ms · 70 in / 27 out · $2.67e-5

FieldExpectedPredictedMatch
nameAmira HassanAmira Hassanmatch
emailamira.hassan@example.comamira.hassan@example.commatch
order_idE-440E-440match
issuelate deliverylate deliverymatch

Judge: field overlap 1.000

{
"name": "Amira Hassan",
"email": "amira.hassan@example.com",
"order_id": "E-440",
"issue": "late delivery"
}
sizeexact 1.000 · schema 1.000 · judge 1.000

Noah Williams (noah.williams@example.com) got the wrong size in order #F-17.

9.8 ms · 66 in / 27 out · $2.61e-5

FieldExpectedPredictedMatch
nameNoah WilliamsNoah Williamsmatch
emailnoah.williams@example.comnoah.williams@example.commatch
order_idF-17F-17match
issuewrong sizewrong sizematch

Judge: field overlap 1.000

{
"name": "Noah Williams",
"email": "noah.williams@example.com",
"order_id": "F-17",
"issue": "wrong size"
}
refundexact 1.000 · schema 1.000 · judge 1.000

Sofia Rossi (sofia.rossi@example.com) has not received the refund for order #G-303.

7.2 ms · 68 in / 28 out · $2.70e-5

FieldExpectedPredictedMatch
nameSofia RossiSofia Rossimatch
emailsofia.rossi@example.comsofia.rossi@example.commatch
order_idG-303G-303match
issuerefund not receivedrefund not receivedmatch

Judge: field overlap 1.000

{
"name": "Sofia Rossi",
"email": "sofia.rossi@example.com",
"order_id": "G-303",
"issue": "refund not received"
}
cancelexact 1.000 · schema 1.000 · judge 1.000

Jonah Kim (jonah.kim@example.com) asked to cancel order #H-88.

6.8 ms · 62 in / 25 out · $2.43e-5

FieldExpectedPredictedMatch
nameJonah KimJonah Kimmatch
emailjonah.kim@example.comjonah.kim@example.commatch
order_idH-88H-88match
issuecancel ordercancel ordermatch

Judge: field overlap 1.000

{
"name": "Jonah Kim",
"email": "jonah.kim@example.com",
"order_id": "H-88",
"issue": "cancel order"
}

fixture-mutate

n 8 · exact 0/8 · schema 4/8

damagedexact 0.000 · schema 1.000 · judge 0.750

Customer Jane Doe (jane@example.com) says order #A-1042 arrived damaged.

8.1 ms · 65 in / 25 out · $2.47e-5

FieldExpectedPredictedMatch
nameJane DoeJane Doematch
emailjane@example.comjane@example.commatch
order_idA-1042a-1042differs
issuearrived damagedarrived damagedmatch

Judge: field overlap 0.750

{
"name": "Jane Doe",
"email": "jane@example.com",
- "order_id": "A-1042",
+ "order_id": "a-1042",
"issue": "arrived damaged"
}
missingexact 0.000 · schema 0.000 · judge 0.750

Luis Ortega (luis.ortega@example.com) cannot find order #B-2201; the item is missing.

8.9 ms · 68 in / 18 out · $2.10e-5

FieldExpectedPredictedMatch
nameLuis OrtegaLuis Ortegamatch
emailluis.ortega@example.com—missing
order_idB-2201B-2201match
issuemissing itemmissing itemmatch
  • email: Field required

Judge: field overlap 0.750

{
"name": "Luis Ortega",
- "email": "luis.ortega@example.com",
"order_id": "B-2201",
"issue": "missing item"
}
billingexact 0.000 · schema 1.000 · judge 0.750

Priya Shah (priya.shah@example.com) was charged twice for order #C-118.

9.6 ms · 65 in / 26 out · $2.53e-5

FieldExpectedPredictedMatch
namePriya ShahPriya Shahmatch
emailpriya.shah@example.compriya.shah@example.commatch
order_idC-118c-118differs
issuecharged twicecharged twicematch

Judge: field overlap 0.750

{
"name": "Priya Shah",
"email": "priya.shah@example.com",
- "order_id": "C-118",
+ "order_id": "c-118",
"issue": "charged twice"
}
addressexact 0.000 · schema 0.000 · judge 0.750

Kenji Sato (kenji.sato@example.com) received order #D-9 at the wrong address.

9.2 ms · 66 in / 17 out · $2.01e-5

FieldExpectedPredictedMatch
nameKenji SatoKenji Satomatch
emailkenji.sato@example.com—missing
order_idD-9D-9match
issuewrong addresswrong addressmatch
  • email: Field required

Judge: field overlap 0.750

{
"name": "Kenji Sato",
- "email": "kenji.sato@example.com",
"order_id": "D-9",
"issue": "wrong address"
}
lateexact 0.000 · schema 1.000 · judge 0.750

Amira Hassan (amira.hassan@example.com) is still waiting on order #E-440; delivery is late.

8.2 ms · 70 in / 27 out · $2.67e-5

FieldExpectedPredictedMatch
nameAmira HassanAmira Hassanmatch
emailamira.hassan@example.comamira.hassan@example.commatch
order_idE-440e-440differs
issuelate deliverylate deliverymatch

Judge: field overlap 0.750

{
"name": "Amira Hassan",
"email": "amira.hassan@example.com",
- "order_id": "E-440",
+ "order_id": "e-440",
"issue": "late delivery"
}
sizeexact 0.000 · schema 0.000 · judge 0.750

Noah Williams (noah.williams@example.com) got the wrong size in order #F-17.

7.8 ms · 66 in / 17 out · $2.01e-5

FieldExpectedPredictedMatch
nameNoah WilliamsNoah Williamsmatch
emailnoah.williams@example.com—missing
order_idF-17F-17match
issuewrong sizewrong sizematch
  • email: Field required

Judge: field overlap 0.750

{
"name": "Noah Williams",
- "email": "noah.williams@example.com",
"order_id": "F-17",
"issue": "wrong size"
}
refundexact 0.000 · schema 1.000 · judge 0.750

Sofia Rossi (sofia.rossi@example.com) has not received the refund for order #G-303.

6.9 ms · 68 in / 28 out · $2.70e-5

FieldExpectedPredictedMatch
nameSofia RossiSofia Rossimatch
emailsofia.rossi@example.comsofia.rossi@example.commatch
order_idG-303g-303differs
issuerefund not receivedrefund not receivedmatch

Judge: field overlap 0.750

{
"name": "Sofia Rossi",
"email": "sofia.rossi@example.com",
- "order_id": "G-303",
+ "order_id": "g-303",
"issue": "refund not received"
}
cancelexact 0.000 · schema 0.000 · judge 0.750

Jonah Kim (jonah.kim@example.com) asked to cancel order #H-88.

7.1 ms · 62 in / 17 out · $1.95e-5

FieldExpectedPredictedMatch
nameJonah KimJonah Kimmatch
emailjonah.kim@example.com—missing
order_idH-88H-88match
issuecancel ordercancel ordermatch
  • email: Field required

Judge: field overlap 0.750

{
"name": "Jonah Kim",
- "email": "jonah.kim@example.com",
"order_id": "H-88",
"issue": "cancel order"
}

fixture-prose

n 8 · exact 0/8 · schema 0/8

damagedexact 0.000 · schema 0.000 · judge 0.000

Customer Jane Doe (jane@example.com) says order #A-1042 arrived damaged.

7.7 ms · 65 in / 15 out · $1.87e-5

FieldExpectedPredictedMatch
nameJane Doe—no JSON
emailjane@example.com—no JSON
order_idA-1042—no JSON
issuearrived damaged—no JSON
  • output is not valid JSON

Judge: field overlap 0.000

- {
- "name": "Jane Doe",
- "email": "jane@example.com",
- "order_id": "A-1042",
- "issue": "arrived damaged"
- }
+ I extracted a ticket but this fixture does not emit JSON.
missingexact 0.000 · schema 0.000 · judge 0.000

Luis Ortega (luis.ortega@example.com) cannot find order #B-2201; the item is missing.

8.3 ms · 68 in / 15 out · $1.92e-5

FieldExpectedPredictedMatch
nameLuis Ortega—no JSON
emailluis.ortega@example.com—no JSON
order_idB-2201—no JSON
issuemissing item—no JSON
  • output is not valid JSON

Judge: field overlap 0.000

- {
- "name": "Luis Ortega",
- "email": "luis.ortega@example.com",
- "order_id": "B-2201",
- "issue": "missing item"
- }
+ I extracted a ticket but this fixture does not emit JSON.
billingexact 0.000 · schema 0.000 · judge 0.000

Priya Shah (priya.shah@example.com) was charged twice for order #C-118.

7.9 ms · 65 in / 15 out · $1.87e-5

FieldExpectedPredictedMatch
namePriya Shah—no JSON
emailpriya.shah@example.com—no JSON
order_idC-118—no JSON
issuecharged twice—no JSON
  • output is not valid JSON

Judge: field overlap 0.000

- {
- "name": "Priya Shah",
- "email": "priya.shah@example.com",
- "order_id": "C-118",
- "issue": "charged twice"
- }
+ I extracted a ticket but this fixture does not emit JSON.
addressexact 0.000 · schema 0.000 · judge 0.000

Kenji Sato (kenji.sato@example.com) received order #D-9 at the wrong address.

6.9 ms · 66 in / 15 out · $1.89e-5

FieldExpectedPredictedMatch
nameKenji Sato—no JSON
emailkenji.sato@example.com—no JSON
order_idD-9—no JSON
issuewrong address—no JSON
  • output is not valid JSON

Judge: field overlap 0.000

- {
- "name": "Kenji Sato",
- "email": "kenji.sato@example.com",
- "order_id": "D-9",
- "issue": "wrong address"
- }
+ I extracted a ticket but this fixture does not emit JSON.
lateexact 0.000 · schema 0.000 · judge 0.000

Amira Hassan (amira.hassan@example.com) is still waiting on order #E-440; delivery is late.

7.1 ms · 70 in / 15 out · $1.95e-5

FieldExpectedPredictedMatch
nameAmira Hassan—no JSON
emailamira.hassan@example.com—no JSON
order_idE-440—no JSON
issuelate delivery—no JSON
  • output is not valid JSON

Judge: field overlap 0.000

- {
- "name": "Amira Hassan",
- "email": "amira.hassan@example.com",
- "order_id": "E-440",
- "issue": "late delivery"
- }
+ I extracted a ticket but this fixture does not emit JSON.
sizeexact 0.000 · schema 0.000 · judge 0.000

Noah Williams (noah.williams@example.com) got the wrong size in order #F-17.

6.9 ms · 66 in / 15 out · $1.89e-5

FieldExpectedPredictedMatch
nameNoah Williams—no JSON
emailnoah.williams@example.com—no JSON
order_idF-17—no JSON
issuewrong size—no JSON
  • output is not valid JSON

Judge: field overlap 0.000

- {
- "name": "Noah Williams",
- "email": "noah.williams@example.com",
- "order_id": "F-17",
- "issue": "wrong size"
- }
+ I extracted a ticket but this fixture does not emit JSON.
refundexact 0.000 · schema 0.000 · judge 0.000

Sofia Rossi (sofia.rossi@example.com) has not received the refund for order #G-303.

8.5 ms · 68 in / 15 out · $1.92e-5

FieldExpectedPredictedMatch
nameSofia Rossi—no JSON
emailsofia.rossi@example.com—no JSON
order_idG-303—no JSON
issuerefund not received—no JSON
  • output is not valid JSON

Judge: field overlap 0.000

- {
- "name": "Sofia Rossi",
- "email": "sofia.rossi@example.com",
- "order_id": "G-303",
- "issue": "refund not received"
- }
+ I extracted a ticket but this fixture does not emit JSON.
cancelexact 0.000 · schema 0.000 · judge 0.000

Jonah Kim (jonah.kim@example.com) asked to cancel order #H-88.

6.2 ms · 62 in / 15 out · $1.83e-5

FieldExpectedPredictedMatch
nameJonah Kim—no JSON
emailjonah.kim@example.com—no JSON
order_idH-88—no JSON
issuecancel order—no JSON
  • output is not valid JSON

Judge: field overlap 0.000

- {
- "name": "Jonah Kim",
- "email": "jonah.kim@example.com",
- "order_id": "H-88",
- "issue": "cancel order"
- }
+ I extracted a ticket but this fixture does not emit JSON.