# Model Capabilities Manifest

**Source of truth**: `scripts/model-capabilities.json`
**Generator**: `scripts/generate-model-capabilities.mjs`
**Emitted artifacts**:
- `crates/buzz-agent/src/generated_model_capabilities.rs`
- `desktop/src/features/agents/ui/modelCapabilities.ts`

## How to regenerate

```sh
node scripts/generate-model-capabilities.mjs
```

CI regenerates and diffs on every PR that touches the manifest, generator, or generated
files. Any stale generated file fails the `model-capabilities` job in `ci.yml`.

## Resolver contract (plan v4)

Resolution is a total function `resolve(provider, raw_model_id) → CapabilityResult`.
Three ordered steps:

1. **Provider-qualified raw exact lookup** — key is `(provider, raw_model_id)`, matched on the
   RAW ID **before any prefix stripping**. A prefixed alias never inherits an exact record.
2. **Provider-scoped ordered family rules** — on the normalized (prefix-stripped) alias.
   Rules ordered by `match_priority` descending (higher wins). Each rule is tagged with the
   providers it applies to.
3. **Per-axis provider fallback** — `blank` (empty model string) vs `concrete_unknown`
   (nonblank but unmatched), per provider.

`CapabilityResult` is complete — every axis is populated. Runtime consumers never compose fields
from multiple tiers.

## Axes (schema fields)

| Axis | Type | Notes |
|------|------|-------|
| `registry_label` | `string \| null` | Optional static display label. Feeds `resolveModelLabel()` registry tier only. |
| `thinking_mode` | enum | `manual-budget \| adaptive \| omit-fields \| none \| not-applicable` |
| `supported_efforts` | `ThinkingEffort[]` | Non-empty. UI effort dropdown options. |
| `default_effort` | `ThinkingEffort \| null` | null = "Inherit" (Anthropic manual-budget models). |
| `databricks_v2_wire_route` | enum | `openai-responses \| anthropic-messages \| mlflow-chat \| route-unknown \| not-applicable` |
| `normalization_policy` | enum | `none \| openai-standard \| openai-clamp-max-to-xhigh` |

### `thinking_mode` values

| Value | Meaning |
|-------|---------|
| `manual-budget` | `thinking:{type:"enabled", budget_tokens}` -- claude-3*, claude-opus-4-5 |
| `adaptive` | `thinking:{type:"adaptive"}` + `output_config:{effort}` -- opus-4-6+, sonnet-4-6+, etc. |
| `omit-fields` | Unknown Anthropic model -- omit thinking fields rather than guess request shape |
| `none` | Non-Anthropic-routed model -- thinking fields not applicable |
| `not-applicable` | Provider does not use Anthropic thinking API |

### `databricks_v2_wire_route` values

Scoped to DBv2 only. All non-DBv2 providers emit `not-applicable`.
Transport for pure OpenAI, legacy Databricks, and OpenRouter is selected by `OpenAiApi` /
`openai_request()` at runtime.

| Value | Meaning |
|-------|---------|
| `openai-responses` | `/ai-gateway/openai/v1/responses` |
| `anthropic-messages` | `/ai-gateway/anthropic/v1/messages` |
| `mlflow-chat` | `/ai-gateway/mlflow/v1/chat/completions` |
| `route-unknown` | DBv2 blank model -- route not yet determinable |
| `not-applicable` | Not a DBv2 provider |

## Family rule match kinds

| Kind | Semantics |
|------|-----------|
| `exact` | Case-insensitive exact string equality on normalized alias |
| `prefix` | Normalized alias starts with match_value |
| `gpt5-token` | Boundary-aware token: present at end-of-string or followed by `-` (not digit/letter) |
| `gpt5-base` | Like gpt5-token but also rejects `-<1-3 digit>` suffixes (version-number rejection) |
| `segment` | Normalized alias contains match_value as a full alphanumeric segment (split on non-alnum) |
| `segment-prefix` | Any segment of the normalized alias starts with match_value |

## Boundaries the manifest does NOT own

- **Transport/endpoint selection for pure OpenAI, legacy Databricks, OpenRouter**: `OpenAiApi` and
  `openai_request()` remain authoritative. The `databricks_v2_wire_route` axis is DBv2-only.
- **Final display labels**: `resolveModelLabel(discovered_name, registry_label, raw_id)` three-tier
  precedence is authoritative. The manifest's `registry_label` feeds only the static registry tier.
- **`llm.rs` replacement scope**: only `databricks_v2_route_for_model`. Other dispatch paths remain.

## Reconciliation policy (plan v4 §Behavior policy)

Not purely behavior-preserving. `models.dev` `reasoning_options` become exact overrides. Each
divergence from family rule results is reconciled against provider docs and either:
- (a) **adopted** as an intentional correction with its own test + exact record, or
- (b) **rejected** with a curation note in the exact record.

### models.dev reconciliation table

**Source queried**: https://models.dev/api.json (2026-07-31)
**Payload SHA-256**: `d5a4974cd69f19b0f67713acaa6bb3b16e920defdc07ecbdf6b0a936181bb0e0`
**Verbatim source snapshot SHA-256** (catalog-sample-fixture.json, deleted in Phase 3):
`dc4092a04392f258bea65de2cef53cb1902dce1779dc2b1b2e21fb56774f2d78`

#### `databricks-gpt-5-4-mini`

| | Family rule (gpt5-4) | models.dev | Disposition |
|---|---|---|---|
| `supported_efforts` | `[none, low, medium, high, xhigh]` | `[low, medium, high]` | **ADOPT** |

Provider-advertised wins per plan F1 policy. Source: providers.databricks.models["databricks-gpt-5-4-mini"].reasoning_options
(retrieved 2026-07-31).

#### `databricks-gpt-5-4-nano`

| | Family rule (gpt5-4) | models.dev | Disposition |
|---|---|---|---|
| `supported_efforts` | `[none, low, medium, high, xhigh]` | `[low, medium, high]` | **ADOPT** |

Same as `databricks-gpt-5-4-mini`.

#### `databricks-gpt-5-6-sol`

| | Family rule (gpt5-6) | models.dev | Disposition |
|---|---|---|---|
| `supported_efforts` | `[none, low, medium, high, xhigh, max]` | `[low, medium, high, max]` | **ADOPT** |

Provider-advertised wins. Source: providers.databricks.models["databricks-gpt-5-6-sol"].reasoning_options
(retrieved 2026-07-31).

#### `databricks-gpt-5-5`

| | Family rule (gpt5-5) | models.dev | Disposition |
|---|---|---|---|
| `supported_efforts` | `[none, low, medium, high, xhigh]` | `[low, medium, high]` | **ADOPT** |

Provider-advertised wins. Source: providers.databricks.models["databricks-gpt-5-5"].reasoning_options
(retrieved 2026-07-31).

#### `databricks-claude-opus-4-7`

| | Family rule | models.dev | Disposition |
|---|---|---|---|
| `reasoning_options` type | effort-based | `budget_tokens` | **NO EFFORT DIVERGENCE** |

models.dev advertises a different capability axis (extended thinking token budget), not an
effort-level selector. No effort divergence to reconcile. Source: providers.databricks.models
["databricks-claude-opus-4-7"].reasoning_options (retrieved 2026-07-31).

### Non-divergences (confirmed consistent)

| Model family | Source | Status |
|---|---|---|
| `claude-opus-4-7`, `claude-opus-4-8` | Anthropic extended-thinking docs (July 2025) | ok |
| `claude-sonnet-5.*`, `claude-fable-5`, `claude-mythos-5` | Anthropic extended-thinking docs (July 2025) | ok |
| `claude-opus-4-6`, `claude-sonnet-4-6`, `claude-mythos-preview` | Anthropic extended-thinking docs (July 2025) | ok |
| `claude-3*` | Anthropic extended-thinking docs (July 2025) | ok |
| `gpt-5-pro`, `gpt-5.6`, `gpt-5.5`, `gpt-5.4`, `gpt-5.1`, `gpt-5` | OpenAI reasoning guide (July 2025) | ok |

## Mutation evidence (historical record)

Mutation testing was run at Phase 2 completion (2026-07-31). 7 generator mutations were applied
in isolation against both TS and Rust interpreters. All 7 were killed by both interpreters (7/7).
The mutation runner (`scripts/run-mutation-evidence.mjs`) was deleted in Phase 3; the normative
corpus (`scripts/normative-corpus.json`) that kills these mutations continues to run in CI.

## Adding a new model family

1. Add a `family_rules` entry with a new unique `id`, appropriate `match_kind`, `providers`,
   `match_priority`, and all capability axes.
2. Run `node scripts/generate-model-capabilities.mjs` to regenerate artifacts.
3. CI verifies byte-clean regeneration.
4. The normative corpus (`scripts/normative-corpus.json`) may need new vectors.

## Adding an exact model override

1. Add an `exact_records` entry with `provider` + `raw_model_id` (the full raw ID, no prefix
   stripping). Include a `_reconciliation` note and doc citation.
2. Run `node scripts/generate-model-capabilities.mjs` -- completeness validator will fail if any
   axis cannot be resolved.
3. Regenerate and commit.