Skip to content

Conversation

@vdimarco
Copy link

Summary

This PR adds the Gatewayz provider to models.dev with 351 AI models from 50+ providers accessible through their unified API gateway.

Provider Details

Models Added (351 total)

Gatewayz aggregates models from multiple AI providers including:

Major Providers:

  • OpenAI (47): GPT-3.5, GPT-4 variants, GPT-4o, GPT-5, o1/o3/o4 reasoning series
  • Qwen (46): 2.5, 3 series, QwQ reasoning, VL multimodal variants
  • Mistral AI (36): Codestral, Devstral, Magistral, Ministral, Mixtral, Pixtral
  • Google (24): Gemini 2.0, 2.5, Gemma 2, 3 variants
  • Meta Llama (21): 3, 3.1, 3.2, 3.3, 4, Guard series
  • DeepSeek (19): R1 reasoning, Chat v3, Prover, distilled versions
  • Anthropic (13): Claude 3, 3.5, 3.7, 4, 4.5, Opus 4
  • Microsoft (9): Phi 3, 4, MAI Data Science R1, WizardLM
  • Moonshotai (7): Kimi K2, Linear hybrid attention models
  • NousResearch (7): Hermes 2, 3, 4 series
  • xAI (7): Grok 3, 4, Grok Code Fast
  • Z-AI (7): GLM 4.5, 4.6 variants
  • NVIDIA (7): Nemotron series
  • Perplexity (6): Sonar, Deep Research, Reasoning
  • Baidu (5): ERNIE 4.5, VL multimodal variants
  • Arcee AI (5): AFM, Coder, Maestro, Spotlight, Virtuoso

And 35+ other providers: AI21, AION Labs, AllenAI, Cohere, DeepCogito, EleutherAI, Inception, Inflection, Liquid AI, MiniMax, Meituan, OpenGVLab, OpenRouter, sao10k, StepFun, Tencent, TheDrummer, TNG Tech, and more.

Files Added

  • providers/gatewayz/provider.toml - Provider configuration
  • providers/gatewayz/models/*.toml - 351 model definition files

Validation

All model definitions follow the models.dev schema with:

  • Model capabilities (attachments, reasoning, tool calling, temperature)
  • Context window and output limits
  • Input/output modalities
  • Release and update dates
  • Open weights status

This makes Gatewayz the largest single provider addition to models.dev, offering comprehensive access to the latest AI models through a unified API.

vdimarco and others added 3 commits November 1, 2025 21:03
Gatewayz is an OpenAI-compatible AI model gateway that aggregates models from 14+ different providers. This configuration includes all 57 currently available models with their respective capabilities, context windows, and modality information.

The provider uses the @ai-sdk/openai-compatible SDK with the API endpoint at https://api.gatewayz.ai/v1.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <[email protected]>
…der-models-txvyj0

Add Gatewayz provider with 58 new models
This commit adds the remaining models from the Gatewayz API, expanding coverage from 57 to 351 models across 50+ providers including:

Major additions:
- Meta Llama: 21 models (3, 3.1, 3.2, 3.3, 4, Guard variants)
- OpenAI: 39 additional models (GPT-3.5, GPT-4 variants, o1, o3, o4 series)
- Mistral AI: 35 additional models (Codestral, Devstral, Magistral, Ministral, Mixtral)
- Google: 21 additional models (Gemini 2.0, 2.5, Gemma 2, 3 variants)
- Qwen: 32 additional models (2.5, 3 series, QwQ)
- DeepSeek: 16 additional models (R1, Chat v3, Prover)
- Anthropic: 11 additional models (Claude 3, 3.5, 3.7, Opus 4 series)
- Microsoft: 9 models (Phi 3, 4, MAI, WizardLM)
- Cohere: 4 models (Command series)
- NousResearch: 7 models (Hermes series)
- Moonshotai: 7 models (Kimi K2, Linear)
- Perplexity: 5 additional models (Sonar variants)

And 30+ other providers including AI21, AION Labs, AllenAI, Arcee AI, Baidu, ByteDance,
Cognitive Computations, DeepCogito, EleutherAI, Inception, Inflection, Liquid AI, NVIDIA,
OpenGVLab, sao10k, StepFun, Tencent, TheDrummer, TNG Tech, xAI, Z-AI, and more.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant