Struct CertificationRow 

pub struct CertificationRow {

    pub model_id: String,
    pub family: String,
    pub parameters: String,
    pub size_category: SizeCategory,
    pub status: ModelStatus,
    pub mqs_score: u32,
    pub grade: String,
    pub certified_tier: String,
    pub last_certified: DateTime<Utc>,
    pub g1: bool,
    pub g2: bool,
    pub g3: bool,
    pub g4: bool,
    pub tps_gguf_cpu: Option<f64>,
    pub tps_gguf_gpu: Option<f64>,
    pub tps_apr_cpu: Option<f64>,
    pub tps_apr_gpu: Option<f64>,
    pub tps_st_cpu: Option<f64>,
    pub tps_st_gpu: Option<f64>,
    pub provenance_verified: bool,
}

A single row from the certification lookup table (models.csv).

This struct represents the complete certification state for a model variant, including MQS score, gateway results, and performance metrics.

The boolean fields (g1-g4, provenance_verified) match the CSV schema and represent gateway pass/fail state directly from test results.

Fields

model_id: String

HuggingFace model ID (e.g., “Qwen/Qwen2.5-Coder-0.5B-Instruct”)

family: String

Model family (e.g., “qwen-coder”, “llama”, “mistral”)

parameters: String

Parameter count string (e.g., “0.5B”, “1.5B”, “7B”)

size_category: SizeCategory

Size category for resource scheduling

status: ModelStatus

Certification status

mqs_score: u32

Model Qualification Score (0-1000)

grade: String

Letter grade (A, B, C, D, F, or “-” for ungraded)

certified_tier: String

Highest certified tier (quick, smoke, mvp, full, or “none”)

last_certified: DateTime<Utc>

Last certification timestamp (ISO8601)

g1: bool

G1: Model loads successfully

g2: bool

G2: Basic inference works

g3: bool

G3: No crashes or panics

g4: bool

G4: Output is not garbage

tps_gguf_cpu: Option<f64>

GGUF format, CPU backend

tps_gguf_gpu: Option<f64>

GGUF format, GPU backend

tps_apr_cpu: Option<f64>

APR format, CPU backend

tps_apr_gpu: Option<f64>

APR format, GPU backend

tps_st_cpu: Option<f64>

SafeTensors format, CPU backend

tps_st_gpu: Option<f64>

SafeTensors format, GPU backend

provenance_verified: bool

Whether model provenance has been verified

Implementations

impl CertificationRow

pub fn new(model_id: impl Into<String>, family: impl Into<String>) -> Self

Create a new certification row for a model.

pub const fn all_gateways_passed(&self) -> bool

Check if all gateway checks passed.

pub fn derive_status(&self) -> ModelStatus

Derive status from MQS score and gateway results.

Follows the specification:

• CERTIFIED: MQS >= 800 AND all gateways passed
• BLOCKED: otherwise

pub fn derive_grade(&self) -> String

Derive grade from MQS score.

Grade thresholds:

• A: 900-1000
• B: 800-899
• C: 600-799
• D: 400-599
• F: 0-399

Trait Implementations

impl Clone for CertificationRow

fn clone(&self) -> CertificationRow

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for CertificationRow

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for CertificationRow

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for CertificationRow

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for CertificationRow

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.