Models

Model IDs available through Floodgate and when to use each one.

Pass one of these exact IDs in the model parameter of your API request or coding agent configuration.

---

OpenAI GPT Series

Model	Model ID	Context	Modality	Best for
GPT-6 Astra	gpt-6-astra	1.1M	Text, Vision	OpenAI’s frontier flagship for complex reasoning, autonomous coding, and difficult long-horizon tasks.
GPT-6 Sol	gpt-6-sol	1.1M	Text, Vision	Balanced frontier model optimized for coding agents (Codex), complex development, and tool use.
GPT-6 Luna	gpt-6-luna	1.1M	Text, Vision	Lightweight, fast, and cost-efficient frontier model for high-volume tasks, quick queries, and summarization.
GPT-5.6-sol	gpt-5.6-sol	1.1M	Text, Vision	Coding agents, refactoring, tool use, and general software engineering.
GPT-5.6-luna	gpt-5.6-luna	1.1M	Text, Vision	Fast, economical GPT-5.6 tier for high-frequency workloads.

---

Image Generation & Editing Models

Model	Model ID	Pricing	Supported Resolutions	Best for
GPT Image 2	gpt-image-2	$0.30 / request	1024x1024, 1536x1024, 1024x1536	Fast, high-throughput image generation and image editing.
GPT Image 2.5	gpt-image-2.5	$0.40 / request	1024x1024, 1536x1024, 1024x1536	Flagship multimodal image generation with superior detail and prompt fidelity.

See the Image Generation Guide for complete request examples, inpainting, and parameters.

---

Anthropic Claude Series

Model	Model ID	Context	Modality	Best for
Claude Opus 5.5	claude-opus-5-5	1M	Text, Vision	Anthropic’s flagship model for deep architectural reasoning, full-repo refactors, and complex agentic tasks.
Claude Opus 5	claude-opus-5	1M	Text, Vision	Top-tier Claude model for Claude Code CLI and challenging engineering problems.
Claude Opus 4.8	claude-opus-4-8	1M	Text, Vision	Ultra-reliable reasoning and long-horizon autonomous agents.
Claude Sonnet 5	claude-sonnet-5	1M	Text, Vision	Fast, balanced intelligence for everyday development, code review, and multimodal analysis.
Claude Fable 5.1	claude-fable-5-1	1M	Text, Vision	Complex analytical tasks, creative generation, and long-horizon workflows.

---

xAI Grok

Model	Model ID	Context	Modality	Best for
Grok 4.7	grok-4.7	1M	Text, Vision	Flagship frontier intelligence for mathematics, deep reasoning, coding, and Grok Build CLI.

---

DeepSeek Series

Model	Model ID	Context	Modality	Best for
DeepSeek V4.1 Flash	deepseek-v4.1-flash	1M	Text, Vision	Next-generation ultra-fast reasoning model with high throughput and low cost.

---

Leading Open Weights & Frontier Models

Model	Model ID	Context	Modality	Best for
Qwen 3.8 Max	qwen3.8-max	1M	Text, Vision	2.4-trillion-parameter MoE flagship for coding, multimodal understanding, and autonomous agents.
Kimi K3	kimi-k3	1M	Text, Vision	Multimodal reasoning model with 1M context and toggleable deep-thinking mode.
GLM-5.3	glm-5.3	1M	Text	Long-horizon coding, agent workflows, and multilingual development.
GLM-5.3 Flash	glm-5.3-flash	1M	Text, Vision	Native multimodal model for quick, efficient coding and everyday agent tasks.

---

Recommended Models by Use Case

Coding Agents

• Codex CLI: gpt-6-sol (Recommended), gpt-6-astra, gpt-5.6-sol
• Claude Code: claude-opus-5-5 (Recommended), claude-opus-5, claude-sonnet-5
• Grok Build: grok-4.7 (Recommended), gpt-6-sol
• OpenCode / Pi / Aider: gpt-6-sol, claude-opus-5-5, deepseek-v4.1-flash, qwen3.8-max

Speed & Efficiency

• Fast everyday coding & reasoning: deepseek-v4.1-flash or glm-5.3-flash
• Affordable high-volume tasks: gpt-6-luna or deepseek-v4.1-flash
• Low latency interactive assistance: claude-sonnet-5

Complex Reasoning & Architecture

• Frontier system architecture & hardest bugs: gpt-6-astra or claude-opus-5-5
• Large-scale multi-file development: gpt-6-sol or claude-opus-5
• Math, algorithmic reasoning, & logic: grok-4.7 or qwen3.8-max

Image Generation

• High-volume generation & quick drafts: gpt-image-2 ($0.30/call)
• Production assets & maximum detail: gpt-image-2.5 ($0.40/call)

---

Discover Available Models via API

To list all models currently enabled for your API key, query the /v1/models endpoint:

curl "https://api.floodgate.dev/v1/models" \
  -H "Authorization: Bearer $FLOODGATE_API_KEY"

Model access is determined by your subscription plan or top-up balance. The /v1/models endpoint always reflects your live account capabilities.