Rename built-in models to the tabby-1-* family

[FuJacob]

Summary

Aligns the four built-in downloadable model presets under a single, version-prefixed naming scheme so future iterations land as a family rather than a one-off rename. Display labels only; the underlying GGUF filenames are untouched.

Before	After
tabby-nano-2	tabby-1-nano
tabby-fast-1	tabby-1-mini
tabby-balanced-1	tabby-1-base
tabby-max-1	tabby-1-pro

Validation

swiftlint lint --quiet                                # exit 0
xcodebuild -project Cotabby.xcodeproj -scheme Cotabby \
  -destination 'platform=macOS' build                 # ** BUILD SUCCEEDED **
xcodebuild test -project Cotabby.xcodeproj -scheme Cotabby \
  -destination 'platform=macOS' CODE_SIGNING_ALLOWED=NO
# Test Suite 'All tests' passed (full suite, 0 failures)

Verified grep -rn for both old and new names: no stale references remain.

Risk / rollout notes

• Pure rename of presentation strings. The RuntimeModelCatalog.displayName(for:) switch is keyed on the GGUF filename, which is unchanged, so already-downloaded models keep working and per-model settings keyed off filename keep resolving correctly.
• Existing users who already have a model installed will see the new label show up immediately on next launch.
• Three files touched: Cotabby/Models/LlamaRuntimeModels.swift (the catalog), CotabbyTests/ModelAndPresentationValueTests.swift (test expectations), and README.md (the user-facing table).

Greptile Summary

This PR renames the four built-in downloadable model presets from ad-hoc names (tabby-nano-2, tabby-fast-1, tabby-balanced-1, tabby-max-1) to a unified tabby-1-* family (tabby-1-nano, tabby-1-mini, tabby-1-base, tabby-1-pro). The change is purely presentational — GGUF filenames used as switch keys and settings identifiers are untouched.

• LlamaRuntimeModels.swift: All four return strings in RuntimeModelCatalog.displayName(for:) updated to the new naming scheme.
• ModelAndPresentationValueTests.swift: Test expectations refreshed for two of the four models; tabby-1-pro and tabby-1-nano assertions are absent.
• README.md: User-facing model table updated to show all four new names alongside their unchanged GGUF filenames and download sources.

Confidence Score: 4/5

Safe to merge — the rename is purely presentational and the underlying GGUF filenames used as persistent identifiers are unchanged.

The catalog logic and README are correct and complete. The test suite only exercises two of the four renamed branches, so a future typo in the tabby-1-pro or tabby-1-nano return strings would go undetected until runtime.

CotabbyTests/ModelAndPresentationValueTests.swift — missing assertions for tabby-1-pro and tabby-1-nano.

Important Files Changed

Filename	Overview
Cotabby/Models/LlamaRuntimeModels.swift	All four display name strings in RuntimeModelCatalog.displayName(for:) correctly renamed to the tabby-1-* family; GGUF filenames (the switch keys) are untouched.
CotabbyTests/ModelAndPresentationValueTests.swift	Test expectations updated for two of the four renamed models; tabby-1-pro and tabby-1-nano assertions are missing, leaving those branches untested.
README.md	Model table updated cleanly to reflect all four new tabby-1-* names; GGUF filenames, sizes, and HuggingFace source links are unchanged.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[GGUF filename\nswitch key] --> B{RuntimeModelCatalog\n.displayName}
    B -->|Qwen3-0.6B-Q4_K_M.gguf| C["tabby-1-mini ✓"]
    B -->|gemma-4-E2B-it-Q4_K_M.gguf| D["tabby-1-base ✓"]
    B -->|gemma-4-E4B-it-Q4_K_M.gguf| E["tabby-1-pro ✓"]
    B -->|SmolLM2-135M-Instruct-q8_0.gguf| F["tabby-1-nano ✓"]
    B -->|anything else| G[raw filename]

    C -->|tested| H([✅ test coverage])
    D -->|tested| H
    E -->|not tested| I([⚠️ no assertion])
    F -->|not tested| I

Loading

_{Reviews (1): Last reviewed commit: "Rename built-in models to the tabby-1-* ..." | Re-trigger Greptile}

Greptile also left 1 inline comment on this PR.

[greptile-apps]

The test function only asserts the new display names for tabby-1-mini and tabby-1-base, leaving tabby-1-pro (gemma-4-E4B-it-Q4_K_M.gguf) and tabby-1-nano (SmolLM2-135M-Instruct-q8_0.gguf) uncovered. A typo in either of those two return strings would not be caught by this suite.

Suggested change

	XCTAssertEqual(
	RuntimeModelCatalog.displayName(for: "Qwen3-0.6B-Q4_K_M.gguf"),
	"tabby-fast-1"
	"tabby-1-mini"
	)
	XCTAssertEqual(
	RuntimeModelCatalog.displayName(for: "gemma-4-E2B-it-Q4_K_M.gguf"),
	"tabby-balanced-1"
	"tabby-1-base"
	)
	// Retired models fall back to their raw filename like any unknown local GGUF.
	XCTAssertEqual(
	RuntimeModelCatalog.displayName(for: "Qwen3-0.6B-Q4_K_M.gguf"),
	"tabby-1-mini"
	)
	XCTAssertEqual(
	RuntimeModelCatalog.displayName(for: "gemma-4-E2B-it-Q4_K_M.gguf"),
	"tabby-1-base"
	)
	XCTAssertEqual(
	RuntimeModelCatalog.displayName(for: "gemma-4-E4B-it-Q4_K_M.gguf"),
	"tabby-1-pro"
	)
	XCTAssertEqual(
	RuntimeModelCatalog.displayName(for: "SmolLM2-135M-Instruct-q8_0.gguf"),
	"tabby-1-nano"
	)
	// Retired models fall back to their raw filename like any unknown local GGUF.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!