Skip to content

chore(cognitive): update AI model catalog#7

Open
github-actions[bot] wants to merge 1 commit intomasterfrom
chore/update-models-11
Open

chore(cognitive): update AI model catalog#7
github-actions[bot] wants to merge 1 commit intomasterfrom
chore/update-models-11

Conversation

@github-actions
Copy link
Copy Markdown
Contributor

Model Update Summary

Updated 26 models across 4 providers (anthropic, groq, google-ai, xai).


Anthropic

New Models Added

Model ID Display Name Lifecycle Input $/1M Output $/1M Context Max Output
claude-opus-4-7 Claude Opus 4.7 production $5 $25 1,000,000 128,000
claude-sonnet-4-6 Claude Sonnet 4.6 production $3 $15 1,000,000 64,000
claude-opus-4-6 Claude Opus 4.6 production $5 $25 1,000,000 128,000
claude-opus-4-5-20251101 Claude Opus 4.5 production $5 $25 200,000 64,000
claude-opus-4-1-20250805 Claude Opus 4.1 deprecated $15 $75 200,000 32,000
claude-opus-4-20250514 Claude Opus 4 deprecated $15 $75 200,000 32,000

Lifecycle Changes

Model ID Change Details
claude-sonnet-4-20250514 productiondeprecated Anthropic announced retirement on 2026-06-15. Added deprecationDate: 2026-06-15, discontinuedDate: 2026-06-15, replacementModels: ['claude-sonnet-4-5-20250929']
claude-3-haiku-20240307 productiondiscontinued Retired by Anthropic on 2026-04-19 (yesterday). Added discontinuedDate: 2026-04-19, replacementModels: ['claude-haiku-4-5-20251001']

Config Changes

  • defaultModel: claude-sonnet-4-5-20250929claude-sonnet-4-6 (updated to current recommended Sonnet)

Groq

Pricing Changes

Model ID Field Old Value New Value
gpt-oss-20b inputCostPer1mTokens $0.10 $0.075
gpt-oss-20b outputCostPer1mTokens $0.50 $0.30
gpt-oss-120b outputCostPer1mTokens $0.75 $0.60

Not Added (insufficient data)

  • groq/compound and groq/compound-mini — compound agentic systems with built-in tools; pricing and output limits not available from docs.
  • Qwen3-32B preview model — listed on Groq console but model ID format and full specs not confirmed.

Google AI

Lifecycle Changes

Model ID Change Details
gemini-3-pro (internalModelId: gemini-3-pro-preview) previewdiscontinued Google shut down Gemini 3 Pro Preview on 2026-03-09. Added discontinuedDate: 2026-03-09, replacementModels: ['gemini-3.1-pro']

Not Added (pricing unavailable from docs)

  • gemini-3.1-pro-preview — confirmed as replacement for gemini-3-pro-preview but no pricing in docs
  • gemini-2.5-flash-lite — new model listed (gemini-2.5-flash-lite) but no pricing available
  • gemini-3.1-flash-lite-preview — new preview model listed but no pricing available
  • gemini-3.1-flash-image-preview, gemini-3.1-flash-live-preview, gemini-3.1-flash-tts-preview — specialized preview models; specs unavailable

xAI

New Models Added

Model ID Display Name Lifecycle Input $/1M Output $/1M Context Max Output
grok-4.20-0309-reasoning Grok 4.20 Reasoning (0309) production $2.00 $6.00 2,000,000 32,768
grok-4.20-0309-non-reasoning Grok 4.20 Non-Reasoning (0309) production $2.00 $6.00 2,000,000 32,768
grok-4.20-multi-agent-0309 Grok 4.20 Multi-Agent (0309) production $2.00 $6.00 2,000,000 128,000
grok-4-1-fast-reasoning Grok 4.1 Fast (Reasoning) production $0.20 $0.50 2,000,000 128,000
grok-4-1-fast-non-reasoning Grok 4.1 Fast (Non-Reasoning) production $0.20 $0.50 2,000,000 128,000

Lifecycle Changes

Model ID Change Replacement
grok-code-fast-1 productiondeprecated grok-4-1-fast-reasoning
grok-4-fast-reasoning productiondeprecated grok-4-1-fast-reasoning
grok-4-fast-non-reasoning productiondeprecated grok-4-1-fast-non-reasoning
grok-4-0709 productiondeprecated grok-4.20-0309-reasoning
grok-3-mini productiondeprecated grok-4-1-fast-non-reasoning
grok-3 productiondeprecated grok-4.20-0309-non-reasoning

Config Changes

  • defaultModel: grok-4-fast-non-reasoninggrok-4-1-fast-non-reasoning

OpenAI

No changes. The OpenAI models documentation page returned 403 Forbidden and could not be fetched. Existing models appear current based on available data.


Cerebras

No changes. The Cerebras introduction page did not expose full model pricing/specs. Existing models appear current.


OpenRouter

No changes. The OpenRouter models page uses dynamic content loading and could not be fully parsed. Existing gpt-oss-120b model remains.


Fireworks AI

No changes. Several potentially new models were spotted on Fireworks AI's website (DeepSeek v3.2, MiniMax M2.7, GLM 5.1, Qwen3.6 Plus, Kimi K2.5) but exact API model IDs and complete specs were not available for confident addition.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant