AI Model Guides
Each AI model has distinct strengths, quirks, and optimal prompting patterns. These guides cover what makes each model unique and how to get the best results from it.
Current models
Specs from each vendor's docs · verified
| Model | Context | Max output | Price / 1M tokens (in / out) | Best for |
|---|---|---|---|---|
GPT-6 AstraOpenAI · gpt-6-astra | 1.05M | 128K | $10 / $50 | Hardest reasoning and long-horizon agentic work |
GPT-6 SolOpenAI · gpt-6-sol | 1.05M | 128K | $2 / $10 | Everyday production workloads |
GPT-6.1 SolOpenAI · gpt-6.1-sol | 1.05M | 128K | $2 / $10 | Everyday production workloads |
GPT-6.1 TerraOpenAI · gpt-6.1-terra | 1.05M | 128K | $1 / $5 | Lower-cost general workloads between Sol and Luna |
GPT-6 LunaOpenAI · gpt-6-luna | 1.05M | 128K | $0.10 / $0.50 | High-volume classification and extraction |
Claude Fable 5.1Anthropic · claude-fable-5-1 | 1M | 128K | $10 / $50 | Anthropic's most capable model for demanding long-running work |
Claude Opus 5.5Anthropic · claude-opus-5-5 | 1M | 128K | $4 / $20 | Coding and agentic work at a lower price than Fable |
Claude Sonnet 5.5Anthropic · claude-sonnet-5-5 | 1M | 128K | $2 / $10 | Daily driver for most production tasks |
Claude Haiku 5.5Anthropic · claude-haiku-5-5 | 1M | 128K | $0.10 / $0.50For prompts up to 100K tokens; $0.50 / $2.50 for prompts over 100K | High-volume classification, extraction and routing |
Gemini 4 ArgonGoogle · gemini-4-argon | 1.05M | 65K | $2 / $10Intro price through Nov 30, 2026; $4 / $20 from Dec 1, 2026 | Google's flagship reasoning model; text output only |
Gemini 3.8 FlashGoogle · gemini-3.8-flash | 1.05M | 65K | $0.75 / $3.75Promo through Dec 31, 2026; $1.50 / $7.50 from Jan 1, 2027 | Fast multimodal work, including native video input |
Gemini 3.5 Flash-LiteGoogle · gemini-3.5-flash-lite | 1.05M | 65K | $0.30 / $2.50 | Cheapest multimodal option for high-volume routes |
Prompting guides
How to Prompt Claude: Fable 5.1, Opus 5.5, Sonnet 5 and Haiku 4.5
Claude responds uniquely well to XML tags, explained constraints and clear output specs. The prompting patterns, thinking and effort settings, and model choices that work best with Claude in 2026.
How to Prompt GPT-6: Astra, Sol and Luna With the Responses API
OpenAI's GPT-6 family (Astra, Sol and Luna) shares a 1.05M-token context window. How to pick a tier, set reasoning effort, get structured outputs, and call tools with the Responses API.
How to Prompt Gemini 3.8 Flash: Video, Long Context, Thinking and Grounding
Gemini 3.8 Flash and 3.5 Flash-Lite take text, images, video, audio and PDFs in a 1M-token window. How to prompt them, set thinking levels, and use Search grounding and code execution.
How to Prompt LLaMA 3: Local Inference and Ollama SetupLegacy
How to run Meta's LLaMA 3 locally with Ollama, write effective prompts for it, and decide when a local model beats a hosted API. Kept for reference; newer open-weight models have since overtaken it.
How to Prompt Mistral: Instruct Format, Efficiency, and API TipsLegacy
Mistral's model family balances strong performance with exceptional efficiency. Learn the instruct format, how to use the Mistral API, and when each model in the family fits best.
GPT-6 vs Claude vs Gemini: Which AI Model for Which Task? (2026)
A practical comparison of GPT-6 Astra, Sol and Luna, Claude Fable 5.1, Opus 5.5 and Sonnet 5, and Gemini 3.8 Flash: context, inputs, reasoning controls and cost for real workloads.
Not sure which model to use?
The comparison guide breaks down OpenAI, Claude, Gemini and open-weight models side by side.
See the comparison