[model-inventory] Model alias inventory update - 2026-07-30

[github-actions]

Warning

threat detection engine error
The threat detection engine encountered an error and could not complete analysis. This is a tooling failure, not a security finding.

Details

The threat detection engine failed to produce results.

Review the workflow run logs for details.

Summary

Analyzed live model inventories from OpenAI, Anthropic, Gemini, Copilot SDK, and Copilot (reflect) against the built-in alias mapping and pricing payload. The alias file (pkg/workflow/data/model_aliases.json) is already well-maintained and covers nearly all current model generations (sonnet-6x, gpt-5.6, gemini-3.x, etc.). One clear gap was found: grok-4.5 (xAI's Grok model, served through Copilot) has no alias coverage and is missing from models.json pricing data entirely, despite being actively served by Copilot with full billing metadata available.

• Providers queried: OpenAI, Anthropic, Gemini, Copilot SDK, Copilot (reflect)
• Total models found: 222 (OpenAI) + 10 (Anthropic) + 50 (Gemini) + 22 (Copilot SDK) + 37 (Copilot reflect)
• Proposed alias changes: 1
• Pricing gaps found: 1 (missing model)

Provider Model Counts

Provider	Models Available	Status
openai	222	✅ ok
anthropic	10	✅ ok
gemini	50	✅ ok
copilot-sdk	22	✅ ok
copilot (reflect)	37	✅ ok

Raw API Fields Discovered

• Copilot SDK / reflect: billing.tokenPrices (inputPrice/outputPrice/cachePrice per-million-token, plus longContext tier pricing), capabilities.limits.max_context_window_tokens, capabilities.supports.vision/tool_calls/reasoning_effort, modelPickerCategory, modelPickerPriceCategory, policy.state.
• Anthropic: display_name, created_at, type — no direct pricing field (matches known family-based pricing tiers already in models.json).
• Gemini: input_token_limit, output_token_limit, supported_generation_methods, version — useful as a complexity/pricing-tier proxy, no direct cost field.
• OpenAI: owned_by, created — used for naming-convention cross-checks only.
• Docs pricing table (https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing): confirms GPT family pricing tiers (Default vs Long context) matching models.json; does not list Anthropic/Gemini/xAI models (rendered separately or omitted from the public table).

models.json Pricing Analysis

Missing from models.json

Model ID	Provider	Inferred Pricing	Basis
grok-4.5	copilot (xAI)	input $0.20/M, output $0.60/M, cache-read $0.05/M (long-context: input $0.40/M, output $1.20/M, cache-read $0.10/M above 200K tokens)	Copilot SDK billing.tokenPrices (inputPrice: 200, outputPrice: 600, cachePrice/cacheReadPrice: 50, longContext.inputPrice: 400, longContext.outputPrice: 1200), corroborated by reflect endpoint model list

Historical entries not currently returned

These model IDs remain in pkg/cli/data/models.json but are absent from all live inventories this run. Retained as historical records — no removal proposed:

claude-fable-5, claude-opus-4.5, claude-sonnet-4, gemini-2.5-pro, gemini-3-flash-preview, gpt-4.1, gpt-5.2, gpt-5.2-codex, gpt-5.4-nano (intentionally deprecated/excluded per policy), kimi-k2.7-code, raptor-mini

Inferred vs stored pricing discrepancies

None found. All GA models present in both models.json and current live inventories (gpt-5.4, gpt-5.5, gpt-5.3-codex, gpt-5-mini, claude-sonnet-4.5, claude-haiku-4.5, etc.) have pricing consistent with the Copilot SDK/reflect tokenPrices and the docs pricing table within rounding tolerance.

Proposed Alias Updates

1. What: Add a new grok alias family
Why: grok-4.5 is actively served by Copilot (confirmed in both Copilot SDK models and the live reflect endpoint model list) but has zero alias coverage today — it only matches the catch-all any alias, with no semantic or family-specific alias, and it's also absent from models.json pricing data.
Syntax:

{
  "grok": ["copilot/*grok*", "openai/*grok*"]
}

Also add grok-4.5 (and its github/grok-4.5 mirror) with the pricing shown above to pkg/cli/data/models.json and actions/setup/js/models.json (keeping the two mirrored files identical), using provider_type: "openai"-style entry consistent with other Copilot-routed non-Anthropic models (needs a maintainer decision on exact provider_type/wire_api since Grok is not currently modeled as its own provider in the schema — the two files today only have anthropic/openai/github-copilot provider buckets).

Full Model Lists by Provider

Anthropic (10): claude-haiku-4-5-20251001, claude-opus-4-1-20250805, claude-opus-4-5-20251101, claude-opus-4-6, claude-opus-4-7, claude-opus-4-8, claude-opus-5, claude-sonnet-4-5-20250929, claude-sonnet-4-6, claude-sonnet-5

Copilot SDK (22): auto, claude-haiku-4.5, claude-opus-4.6, claude-opus-4.7, claude-opus-4.8, claude-opus-5, claude-sonnet-4.5, claude-sonnet-4.6, claude-sonnet-5, gemini-3.1-pro-preview, gemini-3.5-flash, gemini-3.6-flash, gpt-5-mini, gpt-5.3-codex, gpt-5.4, gpt-5.4-mini, gpt-5.5, gpt-5.6-luna, gpt-5.6-sol, gpt-5.6-terra, grok-4.5, mai-code-1-flash-picker

Copilot (reflect, 37): claude-haiku-4.5, claude-opus-4.6/4.7/4.8, claude-sonnet-4.5/4.6/5, gemini-2.5-pro, gemini-3-flash-preview, gpt-3.5-turbo(-0613), gpt-4(-0613), gpt-4-o-preview, gpt-4.1(-2025-04-14), gpt-4o(-2024-05-13/08-06/11-20), gpt-4o-mini(-2024-07-18), gpt-5-mini, gpt-5.3-codex, gpt-5.4(-mini), gpt-5.5, gpt-5.6-luna/sol/terra, grok-4.5, mai-code-1-flash-picker, text-embedding-3-small(-inference), text-embedding-ada-002, trajectory-compaction

OpenAI (222) and Gemini (50) lists omitted for brevity — no new families found requiring alias changes.

Notes

• pkg/cli/data/models.json and actions/setup/js/models.json remain byte-identical (verified via diff) — any pricing update must preserve this parity, e.g. via make refresh-models-json followed by manual addition of the Grok entry (models.dev does not yet list grok-4.5 under GitHub Copilot as of this run).
• gpt-4o-mini, gpt-4.1, gpt-4o, and gpt-5.4-nano continue to be treated as intentionally deprecated per standing policy and are excluded from the gap tables above even though they still appear in the live reflect model list.
• All other alias families (sonnet, sonnet-6x, haiku, opus, gpt-5* variants, gemini-* variants, kimi, mai-code, raptor-mini, task-oriented aliases like summarization, coding, reasoning, vision, image-generation) already have adequate wildcard coverage for all models observed in this run — no changes needed there.

Warning

Firewall blocked 7 domains

The following domains were blocked by the firewall during workflow execution:

• accounts.google.com
• android.clients.google.com
• clients2.google.com
• collector.githubapp.com
• contentautofill.googleapis.com
• safebrowsingohttpgateway.googleapis.com
• www.google.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "accounts.google.com"
    - "android.clients.google.com"
    - "clients2.google.com"
    - "collector.githubapp.com"
    - "contentautofill.googleapis.com"
    - "safebrowsingohttpgateway.googleapis.com"
    - "www.google.com"

See Network Configuration for more information.

Generated by 📦 Daily Model Inventory Checker · aut00 · 50.6 AIC · ⊞ 12.5K · ◷

• expires on Aug 5, 2026, 4:18 PM UTC-08:00