the arsenal
32 models · per-token billing · uncensored tier included on every plan
OpenAI
(8)OpenAI's flagship frontier model. Best for complex reasoning.
Latest GPT-6.1 — top-tier reasoning and coding.
GPT-6 core. Balanced intelligence and speed.
GPT-6 tuned for conversation and creative work.
Proven GPT-5.6 reasoning model, now cheaper.
GPT-5.6 variant with steady, reliable output.
Fast, friendly GPT-5.6 for everyday tasks.
The workhorse. Great quality per credit.
Anthropic
(8)Anthropic's newest creative-writing powerhouse.
The most capable Claude. Elite reasoning, agentic coding.
Previous Opus generation, still exceptional.
The best balance of speed and intelligence.
Proven Sonnet 5 quality at a lower rate.
Refined Opus 4 series. Deep focus work.
Classic Opus muscle.
Value Opus for long sessions.
Z.ai
(6)Frontier coding + reasoning at a tenth of flagship cost. This model built UNJAILED.
GLM-5.3 with safety tuning removed. Fully unfiltered.
Lightning-fast GLM for chat at scale. Cheapest real model on the market.
Uncensored Flash. Fast, cheap, unfiltered.
Abliterated weights — zero refusal behavior.
Previous-gen GLM at a discount.
DeepSeek
(4)Newest DeepSeek. Insane value per token.
DeepSeek V4 Pro with filters stripped.
Flagship DeepSeek reasoning.
Budget DeepSeek for bulk work.
xAI
(3)Latest Grok. Wit, speed, real-time edge.
Grok 4.6 — the internet-aware model.
Discounted Grok generation.
Moonshot AI
(1)Agentic long-context specialist.
Xiaomi
(2)Efficient pro-tier model from Xiaomi.
Ultra-budget flash model.