VVDex ForgeAI exams. Evidence. Independent grading.

Measured capability

Multimodal Annotation

Task-level image, video, text and structured-data annotation evidence, with delivery conditions shown separately.

5 measured exams22 recorded lanesLatest measurement 2026-09-07Published campaign scopeNo category composite

Delivery coverage

Measured and unmeasured stay distinct.

  1. ImageSeparate measurements; no combined rate
    5 conditionsReview measurements ↓
  2. Video — sampled framesCertified passes
    0/10Review measurements ↓
  3. Video — nativeSeparate measurements; no combined rate
    3 conditionsReview measurements ↓
  4. Video — native + exact framesSeparate measurements; no combined rate
    4 conditionsReview measurements ↓
  5. TextSeparate measurements; no combined rate
    31 conditionsReview measurements ↓
  6. Structured dataSeparate measurements; no combined rate
    31 conditionsReview measurements ↓
  7. AudioNo graded attempts
    Not evaluatedCoverage gap

Certified exam

Audio Events 1

1 measured model lane

Model lane

gemini-3.6-flash

Not evaluatedNo graded attempts
Lanegemini-3.6-flashDeliveryDelivery not establishedConditionsUnproven

delivery is not proved

Inspect measurement evidence →

Certified exam

Image object annotation

3 measured model lanes

Model lane · condition 1 of 2

cli/codex-gpt-5.5

2/4Certified passes · rate not published
Lanecli/codex-gpt-5.5DeliveryNative imageConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

cli/codex-gpt-5.5

3/6Certified passes · rate not published
Lanecli/codex-gpt-5.5DeliveryNative imageConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 3

cli/opencode-muse-spark

Not evaluatedNo graded attempts
Lanecli/opencode-muse-sparkDeliveryNative imageConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 3

cli/opencode-muse-spark

0/1Certified passes · rate not published
Lanecli/opencode-muse-sparkDeliveryNative imageConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 3 of 3

cli/opencode-muse-spark

0/100% certified
Lanecli/opencode-muse-sparkDeliveryNative imageConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

gemini-3.6-flash

Not evaluatedNo graded attempts
Lanegemini-3.6-flashDeliveryDelivery not establishedConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Certified exam

Robot video annotation

4 measured model lanes

Model lane · condition 1 of 2

gemini-3.6-flash · gemini-3.6-flash

0/1Certified passes · rate not published
Lanegemini-3.6-flash · gemini-3.6-flashDeliveryVideo — nativeConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

gemini-3.6-flash · gemini-3.6-flash

0/1Certified passes · rate not published
Lanegemini-3.6-flash · gemini-3.6-flashDeliveryVideo — native + exact framesConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 3

meta/muse-spark-1.3 · meta/muse-spark-1.3

1/2Certified passes · rate not published
Lanemeta/muse-spark-1.3 · meta/muse-spark-1.3DeliveryVideo — native + exact framesConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 3

meta/muse-spark-1.3 · meta/muse-spark-1.3

1/1Certified passes · rate not published
Lanemeta/muse-spark-1.3 · meta/muse-spark-1.3DeliveryVideo — native + exact framesConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 3 of 3

meta/muse-spark-1.3 · meta/muse-spark-1.3

3/8Certified passes · rate not published
Lanemeta/muse-spark-1.3 · meta/muse-spark-1.3DeliveryVideo — native + exact framesConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

cli/codex-gpt-5.5

0/9Certified passes · rate not published
Lanecli/codex-gpt-5.5DeliveryVideo — sampled framesConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

cli/codex-gpt-5.5

0/1Certified passes · rate not published
Lanecli/codex-gpt-5.5DeliveryVideo — sampled framesConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 3

gemini-3.8-flash · gemini-3.8-flash

Not evaluatedNo graded attempts
Lanegemini-3.8-flash · gemini-3.8-flashDeliveryVideo — nativeConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 3

gemini-3.8-flash

Not evaluatedNo graded attempts
Lanegemini-3.8-flashDeliveryDelivery not establishedConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 3 of 3

gemini-3.8-flash · gemini-3.8-flash

Not evaluatedNo graded attempts
Lanegemini-3.8-flash · gemini-3.8-flashDeliveryVideo — nativeConditionsUnproven

delivery adapter identity, version or configuration is not recorded; delivery conditions differ; delivery equivalence is unproven; delivery is not proved; different exam fingerprint; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Certified exam

Structured-data labeling

20 measured model lanes

Model lane · condition 1 of 2

@cf/openai/gpt-oss-20b

1/1Certified passes · rate not published
Lane@cf/openai/gpt-oss-20bDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

@cf/openai/gpt-oss-20b

9/1090% certified
Lane@cf/openai/gpt-oss-20bDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

cli/claude-sonnet

1/1Certified passes · rate not published
Lanecli/claude-sonnetDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

cli/claude-sonnet

10/10100% certified
Lanecli/claude-sonnetDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

cli/codex

1/1Certified passes · rate not published
Lanecli/codexDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

cli/codex

10/10100% certified
Lanecli/codexDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

cli/codex-gpt-5.5

10/10100% certified
Lanecli/codex-gpt-5.5DeliveryDelivery not establishedConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

cli/cursor-grok

Not evaluatedNo graded attempts
Lanecli/cursor-grokDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 3

cli/opencode-muse-spark

1/1Certified passes · rate not published
Lanecli/opencode-muse-sparkDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 3

cli/opencode-muse-spark

10/10100% certified
Lanecli/opencode-muse-sparkDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 3 of 3

cli/opencode-muse-spark

Not evaluatedNo graded attempts
Lanecli/opencode-muse-sparkDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

codestral-latest

1/1Certified passes · rate not published
Lanecodestral-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

codestral-latest

9/1090% certified
Lanecodestral-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

cohere/north-mini-code:free

Not evaluatedNo graded attempts
Lanecohere/north-mini-code:freeDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

gemini-3.6-flash

Not evaluatedNo graded attempts
Lanegemini-3.6-flashDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

magistral-small-latest

Not evaluatedNo graded attempts
Lanemagistral-small-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

minimax/minimax-m3:free

1/1Certified passes · rate not published
Laneminimax/minimax-m3:freeDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

ministral-14b-latest

0/1Certified passes · rate not published
Laneministral-14b-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

ministral-14b-latest

8/1080% certified
Laneministral-14b-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

ministral-8b-latest

0/1Certified passes · rate not published
Laneministral-8b-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

ministral-8b-latest

1/1010% certified
Laneministral-8b-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

mistral-code-latest

1/1Certified passes · rate not published
Lanemistral-code-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

mistral-code-latest

9/1090% certified
Lanemistral-code-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

mistral-medium-latest

Not evaluatedNo graded attempts
Lanemistral-medium-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

mistral-small-latest

Not evaluatedNo graded attempts
Lanemistral-small-latestDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free

Not evaluatedNo graded attempts
Lanenvidia/nemotron-3-nano-omni-30b-a3b-reasoning:freeDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

nvidia/nemotron-3-super-120b-a12b:free

Not evaluatedNo graded attempts
Lanenvidia/nemotron-3-super-120b-a12b:freeDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

openai/gpt-oss-120b

1/1Certified passes · rate not published
Laneopenai/gpt-oss-120bDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

openai/gpt-oss-120b

7/1070% certified
Laneopenai/gpt-oss-120bDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

openai/gpt-oss-20b

1/1Certified passes · rate not published
Laneopenai/gpt-oss-20bDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

openai/gpt-oss-20b

10/10100% certified
Laneopenai/gpt-oss-20bDeliveryStructured dataConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Certified exam

Text labeling

20 measured model lanes

Model lane · condition 1 of 2

@cf/openai/gpt-oss-20b

1/1Certified passes · rate not published
Lane@cf/openai/gpt-oss-20bDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

@cf/openai/gpt-oss-20b

10/10100% certified
Lane@cf/openai/gpt-oss-20bDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

cli/claude-sonnet

1/1Certified passes · rate not published
Lanecli/claude-sonnetDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

cli/claude-sonnet

10/10100% certified
Lanecli/claude-sonnetDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

cli/codex

1/1Certified passes · rate not published
Lanecli/codexDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

cli/codex

10/10100% certified
Lanecli/codexDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

cli/codex-gpt-5.5

10/10100% certified
Lanecli/codex-gpt-5.5DeliveryDelivery not establishedConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

cli/cursor-grok

Not evaluatedNo graded attempts
Lanecli/cursor-grokDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 3

cli/opencode-muse-spark

Not evaluatedNo graded attempts
Lanecli/opencode-muse-sparkDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 3

cli/opencode-muse-spark

1/1Certified passes · rate not published
Lanecli/opencode-muse-sparkDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 3 of 3

cli/opencode-muse-spark

10/10100% certified
Lanecli/opencode-muse-sparkDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; recorded engine or harness conditions differ; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

codestral-latest

0/1Certified passes · rate not published
Lanecodestral-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

codestral-latest

0/100% certified
Lanecodestral-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

cohere/north-mini-code:free

Not evaluatedNo graded attempts
Lanecohere/north-mini-code:freeDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

gemini-3.6-flash

Not evaluatedNo graded attempts
Lanegemini-3.6-flashDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

magistral-small-latest

Not evaluatedNo graded attempts
Lanemagistral-small-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

minimax/minimax-m3:free

0/1Certified passes · rate not published
Laneminimax/minimax-m3:freeDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

ministral-14b-latest

0/1Certified passes · rate not published
Laneministral-14b-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

ministral-14b-latest

0/100% certified
Laneministral-14b-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

ministral-8b-latest

0/1Certified passes · rate not published
Laneministral-8b-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

ministral-8b-latest

0/100% certified
Laneministral-8b-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

mistral-code-latest

0/1Certified passes · rate not published
Lanemistral-code-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

mistral-code-latest

0/100% certified
Lanemistral-code-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

mistral-medium-latest

Not evaluatedNo graded attempts
Lanemistral-medium-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

mistral-small-latest

Not evaluatedNo graded attempts
Lanemistral-small-latestDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free

Not evaluatedNo graded attempts
Lanenvidia/nemotron-3-nano-omni-30b-a3b-reasoning:freeDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane

nvidia/nemotron-3-super-120b-a12b:free

1/1Certified passes · rate not published
Lanenvidia/nemotron-3-super-120b-a12b:freeDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

openai/gpt-oss-120b

1/1Certified passes · rate not published
Laneopenai/gpt-oss-120bDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

openai/gpt-oss-120b

3/3Certified passes · rate not published
Laneopenai/gpt-oss-120bDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 1 of 2

openai/gpt-oss-20b

Not evaluatedNo graded attempts
Laneopenai/gpt-oss-20bDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →

Model lane · condition 2 of 2

openai/gpt-oss-20b

1/1Certified passes · rate not published
Laneopenai/gpt-oss-20bDeliveryTextConditionsUnproven

delivery conditions differ; delivery equivalence is unproven; delivery is not proved; lane configuration is not recorded; repository cleanliness differs; execution equivalence is unproven; runtime limits differ

Inspect measurement evidence →