[model-inventory] Model alias inventory update - 2026-08-10

[github-actions]

Summary

Daily model inventory check found the alias map and pricing table to be largely up to date. One gap was identified: a new Copilot model (kimi-k3) is listed in the models.dev catalog but not yet present in models.json, and it is not yet visible in the live Copilot reflect/SDK endpoints for this run (likely not yet rolled out to this account/session).

• Providers queried: OpenAI, Anthropic, Gemini, Copilot SDK, Copilot (reflect)
• Total models found: 222 (openai) + 9 (anthropic) + 50 (gemini) + 22 (copilot-sdk) + 35 (copilot reflect)
• Proposed alias changes: 0 (existing kimi alias pattern copilot/kimi* already covers kimi-k3 once it's added)
• Pricing gaps found: 1 (missing model, not a mismatch)

Provider Model Counts

Provider	Models Available	Status
openai	222	✅ ok
anthropic	9	✅ ok
gemini	50	✅ ok
copilot-sdk	22	✅ ok
copilot (reflect)	35	✅ ok

Raw API Fields Discovered

• Anthropic: display_name, created_at, max_input_tokens, max_tokens, capabilities.* (batch, citations, code_execution, context_management) — useful for capability-aware aliasing.
• Gemini: inputTokenLimit, outputTokenLimit, supportedGenerationMethods, thinking (reasoning flag) — good proxy for pricing tier and vision/tool alias coverage.
• Copilot reflect: pricing.default / pricing.long_context (input/output/cachedInput/cacheWrite/threshold) — matches models.json cost fields exactly for all cross-checked models (claude-sonnet-5, gpt-5.6-sol, grok-4.5).
• Copilot SDK: billing.tokenPrices (inputPrice/outputPrice/cachePrice/cacheWritePrice/contextMax, plus nested longContext) — confirmed consistent with models.json cost fields (values are per-100M-token units, i.e. models.json cost = tokenPrice / 1e8).
• models.dev: normalized cost, limit.context/limit.output, reasoning, modalities.input/output, family, release_date — used here to surface kimi-k3, a model not yet visible in live Copilot endpoints.

models.json Pricing Analysis

Missing from models.json

Model ID	Provider	Inferred Pricing	Basis
kimi-k3	copilot (via models.dev)	input $3.00 / output $15.00 / cache_read $0.30 per 1M tokens	models.dev github-copilot.models["kimi-k3"].cost — not yet present in live reflect or Copilot SDK data for this run

Note: kimi-k3 is not currently returned by the live Copilot reflect endpoint or the Copilot SDK models list, so it may not yet be enabled for this account. No pricing mismatch was found for any currently-live model — all Copilot SDK billing.tokenPrices and reflect pricing.default values cross-checked exactly against the stored cost fields in models.json for claude-sonnet-5, claude-opus-4.6/4.7/4.8/5, claude-haiku-4.5, claude-sonnet-4.5/4.6, gpt-5.4, gpt-5.4-mini, gpt-5.5, gpt-5.6-luna/sol/terra, gpt-5.3-codex, gpt-5-mini, mai-code-1-flash-picker, gemini-3.1-pro-preview, gemini-3.5-flash, gemini-3.6-flash, and grok-4.5.

Historical entries not currently returned

The following models.json (github-copilot) entries are not present in the live reflect/SDK inventory for this run. These appear to be prior-generation or not-yet-rolled-out models; retaining them as historical records per policy:

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

Deprecated Copilot IDs (gpt-4o-mini, gpt-4.1, gpt-4o, gpt-5.4-nano) were seen in the live reflect endpoint but are intentionally excluded from missing/discrepancy tables per the standing exclusion policy.

Inferred vs stored pricing discrepancies

None found. All cross-checked live pricing (Copilot SDK + reflect) matches models.json exactly.

Proposed Alias Updates

None required this run. The existing kimi alias (["copilot/kimi*", "openai/kimi*"]) already uses a wildcard pattern that will automatically match kimi-k3 once it is added to models.json and becomes live — no alias JSON changes are needed.

Recommendation: once kimi-k3 becomes available in the live Copilot inventory (visible in reflect/SDK), run make refresh-models-json (or manually add the entry) using the models.dev pricing above as the seed, then re-verify against live SDK/reflect data at that time.

Full Model Lists by Provider

Copilot (reflect, live, 35 models): 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, 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-41-copilot, 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, grok-4.5, mai-code-1-flash-picker, text-embedding-3-small, text-embedding-3-small-inference, text-embedding-ada-002, trajectory-compaction

Copilot SDK (22 models): 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

Anthropic (9 models): claude-haiku-4-5-20251001, 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

GitHub Docs pricing table (14 rows, OpenAI-family models with GA tiering): GPT-5 mini, GPT-5.3-Codex, GPT-5.4 (default + long context), GPT-5.4 mini, GPT-5.4 nano (excluded per policy), GPT-5.5 (default + long context), GPT-5.6 Luna (default + long context), GPT-5.6 Sol (default + long context), GPT-5.6 Terra (default + long context)

Notes

• The Copilot reflect endpoint reports OpenAI and Anthropic direct-API providers as configured: false in this run (only Copilot's own proxy endpoint is configured), so those cross-checks rely on the separately-fetched provider API keys used for the OpenAI/Anthropic/Gemini inventory steps.
• kimi-k2.7-code remains in models.json as a historical entry; kimi-k3 appears to be its successor per models.dev's family grouping (kimi-k2 → kimi-k3), but is not yet live.
• Existing alias coverage (sonnet, sonnet-6x, haiku, opus, gpt-5.x, gemini-3.x, grok, mai-code, kimi) already matches all currently-live model IDs via wildcard globs — no stale patterns detected.

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 · auto · 56.4 AIC · ⌖ 4.53 AIC · ⊞ 12.7K · ◷

• expires on Aug 16, 2026, 4:09 PM UTC-08:00