Choose Your Model. Maximize Your Credits.

Compare AI models and see how many chats you get per credit.

Credits Breakdown

Not every task needs the most powerful AI. Some need speed. Others need volume. That’s why we offer a range of models — so you choose the right brain for the job and get the most from every credit. This table represents the average credits used per chat message for each model.

Model	Credits	Best For
Claude Opus 4.6	2.30	Top-tier Claude reasoning for complex strategy, nuanced writing, and high-stakes analysis.
Perplexity Sonar Pro	1.70	Fast, citation-friendly answers with live web search for research and fact-checking.
Claude Sonnet 4.5	1.40	Advanced Claude balance of intelligence and speed for demanding day-to-day work.
Claude 3.7 Sonnet	1.20	Deep reasoning for long documents, planning, and accuracy-critical tasks.
Claude 3.5 Sonnet	1.20	Reliable general-purpose model for writing, synthesis, and analysis.
Grok 3	1.20	Strong reasoning and coding support with web-aware responses and quick exploration.
Claude Sonnet 4	1.20	High-capability Claude for structured workflows, summaries, and enterprise use cases.
Perplexity Sonar	0.80	Lightweight web-connected assistant for briefs, quick Q&A, and everyday drafting.
OpenAI GPT-4.1	0.80	Precise instruction following for complex writing, analysis, and coding tasks.
OpenAI GPT-4o	0.80	Fast multimodal GPT for chat, vision-enabled tasks, and broad assistant workflows.
Google Gemini 3.1 Pro Preview	0.80	Preview Gemini model for advanced multimodal reasoning and higher-complexity workflows.
Google Gemini 2.5 Pro	0.60	Multimodal reasoning across text, code, and images for research and content creation.
DeepSeek-R1	0.50	Reasoning-focused open model for technical problem-solving, math, and code explanation.
OpenAI o4-mini	0.40	Cost-efficient OpenAI reasoning model for drafts, prototyping, and lightweight coding.
Claude 4.5 Haiku	0.40	Ultra-fast Claude for short responses, quick edits, and real-time interactions.
Claude 3.5 Haiku	0.30	Speed-first Claude option for summaries, rewrites, and high-throughput tasks.
Amazon Nova Pro	0.30	Balanced performance and cost for internal tools and medium-complexity workflows.
Google Gemini 3 Flash Preview	0.20	Fast preview Gemini model for lightweight prompts, quick drafts, and rapid responses.
DeepSeek 3.2v	0.20	Efficient general model for everyday chat, drafting, and practical coding assistance.
Meta Llama 3.3 70B	0.20	Open model with strong value for creative writing, coding, and experimentation.
Grok 3 Mini	0.10	Low-latency Grok for fast replies, ideation, and lightweight assistant tasks.
OpenAI gpt-oss 120b*	0.00	Large open-weights OpenAI model for flexible experimentation and custom workflows.
OpenAI gpt-oss 20b*	0.00	Small open-weights OpenAI model optimized for speed and low-cost iteration.
Amazon Nova Light*	0.00	Best for high-volume Q&A, lookups, and bulk automation where speed matters most.

Note: Chat counts are approximate and may vary based on usage patterns and prompt complexity.

* these models have a non-zero value, but are below $0.01 per typical chat message.

What's the Best Option for Me?

Here’s a quick guide:

Need the best thinking available? Go with Claude 3.7 Sonnet.
Want something fast and efficient? Claude 3.5 Haiku or DeepSeek-R1 are great picks.
Building tools or experimenting? Use Llama 3.3 70B – amazing value and open-source.
Running high-volume chat or Q&A? Nova Light is unbeatable for scale.

Still Not Sure?

We’ve built this so you’re in control. No mystery pricing. Just clear value per credit.

Tip: For most users, mixing models gives the best balance of quality and cost.

Need help choosing? Contact Us