[model-inventory] Model alias inventory update - 2026-07-28

[github-actions]

Summary

Ran the daily model inventory check. Live provider data (OpenAI, Anthropic, Gemini, Copilot reflect) was
compared against pkg/workflow/data/model_aliases.json and pkg/cli/data/models.json /
actions/setup/js/models.json. All live models from Copilot, Anthropic, and Gemini already resolve
under existing alias patterns — no alias changes were needed. However, models.json was stale versus
the current models.dev catalog and Copilot pricing docs, so it was refreshed and corrected.

• Providers queried: OpenAI, Anthropic, Gemini, Copilot (reflect); Copilot API endpoint returned
  HTTP 400 and was treated as unavailable.
• Total live models found: OpenAI 219, Anthropic 10, Gemini 50, Copilot (reflect) 36
• Proposed alias changes: 0 (all live models already match existing glob patterns)
• Pricing fixes applied directly: models.json refreshed from models.dev/catalog.json, plus manual
  restoration of provider_type/wire_api metadata (dropped by the refresh script), addition of the
  missing gemini-3.6-flash entry, and restoration of raptor-mini (still GA per Copilot pricing docs
  and still referenced by the raptor-mini alias, but absent from the models.dev catalog).

Provider Model Counts

Provider	Models Available	Status
openai	219	✅ ok
anthropic	10	✅ ok
gemini	50	✅ ok
copilot (reflect)	36	✅ ok
copilot-api	0	⚠️ HTTP 400, skipped

Raw API Fields Discovered

Fields worth noting

• OpenAI (raw.json): id, owned_by, created only — no capability/cost fields; naming
  conventions (-mini, -nano, o1/o3/o4) used as the only pricing-tier heuristic.
• Anthropic (raw.json): id, display_name, created_at, type — family grouping via
  display_name (haiku/sonnet/opus) is reliable and used for validation.
• Gemini (raw.json): input_token_limit, output_token_limit, supported_generation_methods,
  version, description — token limits are a useful (if approximate) tier-size proxy.
• Copilot reflect (reflect.json): no billing.multiplier field is present in this run's reflect
  payload (only a flat models id list per endpoint plus a separate model_aliases object mirroring
  the built-in alias file) — billing multiplier validation for Copilot fell back to the docs pricing
  table instead.
• Copilot pricing docs (scraped via Playwright): Model, Release status, Category, Tier,
  Threshold (input tokens), Input, Cached input, Cache write (Anthropic table only), Output —
  most complete and currently-accurate source for Copilot-facing multipliers/pricing tiers.

models.json Pricing Analysis

Missing from models.json (now added)

Model ID	Provider	Fix
gemini-3.6-flash	copilot (github-copilot)	Added: input $1.50 / output $7.50 per docs pricing table, provider_type: openai, wire_api: completions (matched to sibling gemini-3.5-flash entry pattern)

Historical entries not currently returned

The following github-copilot model IDs are absent from the live Copilot reflect endpoint but were
kept in models.json as historical/deprecated entries (some still GA per docs, others long
deprecated) — no automatic removals proposed:

gpt-3.5-turbo, gpt-3.5-turbo-0613, gpt-4, gpt-4-0613, gpt-4-o-preview,
gpt-4.1-2025-04-14, gpt-4o, gpt-4o-2024-05-13, gpt-4o-2024-08-06, gpt-4o-2024-11-20,
gpt-4o-mini, gpt-4o-mini-2024-07-18, text-embedding-3-small,
text-embedding-3-small-inference, text-embedding-ada-002, trajectory-compaction.

Per instructions, gpt-4o-mini, gpt-4.1, gpt-4o, and gpt-5.4-nano are treated as intentionally
deprecated and excluded from missing/discrepancy tables even though gpt-4.1 and gpt-5.4-nano
still appear in models.json/reflect data.

raptor-mini was present in the previous models.json but dropped by models.dev/catalog.json
(the refresh source) — it is still listed as GA in the Copilot pricing docs and still referenced by
the raptor-mini alias, so it was manually restored rather than deleted.

Inferred vs stored pricing discrepancies

No discrepancies found. Cross-checking models.dev-refreshed values against the scraped Copilot
pricing docs table showed input/output pricing matching for every model present in both sources
(Claude Sonnet/Opus/Haiku/Fable family, GPT-5.x family, Gemini 2.5/3.x family, Raptor mini,
MAI-Code-1-Flash, Kimi K2.7 Code).

Fixes Applied

• Ran make refresh-models-json to sync pkg/cli/data/models.json and actions/setup/js/models.json
  from models.dev/catalog.json.
• Restored provider_type / wire_api fields for github-copilot entries (the refresh script strips
  these; they were re-applied from the prior file version where the model ID matched).
• Added gemini-3.6-flash (present in the alias file's pattern and in the Copilot pricing docs table,
  but missing from the github-copilot models list in models.json).
• Restored raptor-mini (dropped by the refresh source but still GA/still aliased).
• Verified make validate-models-json-sync passes (mirrors identical).

Proposed Alias Updates

None. All live Copilot, Anthropic, and Gemini model IDs already resolve against existing glob patterns
in pkg/workflow/data/model_aliases.json (e.g. sonnet-6x, opus, haiku, gpt-5.*, gemini-*-flash,
gemini-*-pro). No new model families or task-oriented alias gaps were identified this run.

Full Model Lists by Provider

Copilot (reflect, 36 models): claude-haiku-4.5, claude-opus-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-2.5-pro,
gemini-3-flash-preview, gpt-3.5-turbo, gpt-3.5-turbo-0613, gpt-4, gpt-4-0613, gpt-4-o-preview, gpt-4.1,
gpt-4.1-2025-04-14, gpt-4o, gpt-4o-2024-05-13, gpt-4o-2024-08-06, gpt-4o-2024-11-20, gpt-4o-mini,
gpt-4o-mini-2024-07-18, 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, mai-code-1-flash-picker, text-embedding-3-small,
text-embedding-3-small-inference, text-embedding-ada-002, trajectory-compaction

Anthropic (10 models): 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

Gemini (50 models, notable): gemini-2.5-flash(-lite/-image), gemini-2.5-pro, gemini-3-flash-preview,
gemini-3-pro-preview, gemini-3.1-flash-lite(-preview), gemini-3.1-pro-preview, gemini-3.5-flash(-lite),
gemini-3.6-flash, gemini-embedding-001/2, gemma-4-26b/-31b, imagen-4.0-, veo-3.1-, lyria-3-,
nano-banana-pro-preview, deep-research--preview

Notes

• Copilot API endpoint (/tmp/.../copilot-api-models/raw.json) returned {"error":"HTTP 400"} and
  was skipped as a validation source this run.
• The Copilot /reflect payload's own model_aliases field mirrors the current alias file exactly,
  confirming the built-in alias set is already in sync with what the live Copilot proxy expects.
• Copilot pricing docs (scraped via Playwright) explicitly exclude gpt-4o-mini, gpt-4.1, gpt-4o,
  and gpt-5.4-nano from the current pricing table — consistent with treating them as deprecated
  per instructions.

Warning

Firewall blocked 8 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
• proxy.golang.org
• 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"
    - "proxy.golang.org"
    - "safebrowsingohttpgateway.googleapis.com"
    - "www.google.com"

See Network Configuration for more information.

Generated by 📦 Daily Model Inventory Checker · sonnet50 · 126.3 AIC · ⌖ 11.5 AIC · ⊞ 12.3K · ◷

• expires on Aug 3, 2026, 8:03 PM UTC-08:00

[github-actions]

This issue is being closed as outdated. A newer issue has been created: #48616

View newer issue

---

This action was performed automatically by the Daily Model Inventory Checker workflow.