[model-inventory] Model alias inventory update - 2026-06-24

[github-actions]

Summary

Daily model alias inventory check (2026-06-24). Four new alias families identified,
one new model missing from models.json (found in models-dev catalog).

• Providers queried: OpenAI (214), Anthropic (8), Gemini (50), Copilot (30 via reflect)
• Proposed alias changes: 4 new aliases (fable, veo, tts, lyria)
• Pricing gaps: 1 — claude-fable-5 missing from github-copilot section of models.json

make refresh-models-json returned HTTP 403 (models.dev unreachable). Findings use the
pre-fetched models-dev/api.json artifact and live provider API responses.

Provider Model Counts

Provider	Models	Status
openai	214	✅ ok
anthropic	8	✅ ok
gemini	50	✅ ok
copilot	30	✅ ok (reflect only; no billing.multiplier in this run)

Raw API Fields Discovered

Anthropic — richest metadata: display_name, created_at, max_input_tokens, max_tokens,
plus capability flags: effort (low/medium/high/xhigh/max), thinking (adaptive), image_input,
code_execution, batch, structured_outputs. All claude-4.x models support effort + thinking.

OpenAI — minimal: id, owned_by, created. No context-window or pricing fields.

Gemini — displayName, inputTokenLimit, outputTokenLimit, supportedGenerationMethods.
TTS models have small input (8 192) / large output (16 384) token profiles vs. standard 1M/64K.

Copilot (reflect) — flat model-ID list only; no billing.multiplier metadata this run.

models.json Pricing Analysis

Missing from models.json

Model ID	Provider	Pricing ($/M tokens)	Source
claude-fable-5	github-copilot	in: $10, out: $50, cache_read: $1, cache_write: $12.5	models-dev

New premium family (release 2026-06-09), 1M-context, reasoning+image support.
Proposed models.json entry (per-token scientific notation):

"claude-fable-5": {
  "cost": {
    "input": "1e-05",
    "output": "5e-05",
    "cache_read": "1e-06",
    "cache_write": "1.25e-05"
  }
}

Historical entries not in live APIs (retain)

github-copilot: claude-sonnet-4, gemini-3.1-pro-preview, gemini-3.5-flash,
gpt-5.2, gpt-5.2-codex, raptor-mini

anthropic: claude-3.x series, claude-opus-4-0/4-1, claude-sonnet-4-0/4-5 (short IDs)

Pricing discrepancies

None detected — all stored entries align with models-dev data.

Proposed Alias Updates

1. fable — New Claude Fable family

claude-fable-5 in models-dev catalog (not yet in Copilot reflect or Anthropic API). Premium
tier above opus ($10/$50 vs $5/$25/M). Needs alias for explicit targeting.

"fable": ["copilot/*fable*", "anthropic/*fable*"]

2. veo — Google Veo video generation

Live Gemini API: veo-2.0-generate-001, veo-3.0-generate-001, veo-3.0-fast-generate-001,
veo-3.1-generate-preview, veo-3.1-fast-generate-preview, veo-3.1-lite-generate-preview.
No existing alias covers this family.

"veo": ["copilot/veo*", "google/veo*", "gemini/veo*"]

3. tts — Text-to-speech (Gemini TTS family)

Live Gemini API: gemini-2.5-flash-preview-tts ($0.5/$10/M), gemini-2.5-pro-preview-tts
($1/$20/M), gemini-3.1-flash-tts-preview. Distinct token profile; no alias exists.

"tts": ["copilot/gemini-*tts*", "copilot/*tts*", "google/gemini-*tts*", "gemini/gemini-*tts*"]

4. lyria — Google Lyria audio/music generation

Live Gemini API: lyria-3-clip-preview, lyria-3-pro-preview. No alias exists.

"lyria": ["copilot/lyria*", "google/lyria*", "gemini/lyria*"]

Notes

• fable intentionally not added to large/agent chains pending confirmation it is live in
  Copilot.
• mai-code alias has no live matches; kept as forward-looking.
• Copilot billing docs scrape returned empty models array — the HTML extractor may need
  updating for the current page layout.
• All existing aliases (sonnet, haiku, opus, gpt-5.*, gemini-3.*, gemini-flash-lite,
  image-generation, deep-research, computer-use, robotics, antigravity,
  nano-banana) continue to cover their respective live model families with no gaps.

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• models.dev

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

network:
  allowed:
    - defaults
    - "models.dev"

See Network Configuration for more information.

Generated by 📦 Daily Model Inventory Checker · 116.1 AIC · ⌖ 15.4 AIC · ⊞ 7.5K · ◷

• expires on Jun 30, 2026, 4:24 PM UTC-08:00