Skip to main content
PRYSM routes across 21 models from 9 providers (OpenAI, Anthropic, Google, DeepSeek, xAI, Mistral, Moonshot, Alibaba, Perplexity). With model: "auto", PRYSM picks the best-value model for each prompt; pass any id below to pin a single request to that model.
Prices are USD per million tokens (MTok). This page mirrors the live catalog — fetch the current data any time from GET /v1/models or prysm models.

​
Budget

Fast and inexpensive — ideal for classification, simple Q&A, and high-volume work.
IDModelProviderContextInputOutputStrengths
mistral-nemoMistral Nemomistral128K$0.02$0.02classification, simple, fast
gpt-5-nanoGPT-5 Nanoopenai128K$0.05$0.40simple, chat, fast
gemini-2.5-flash-liteGemini 2.5 Flash-Litegoogle1M$0.10$0.40simple, translate, multimodal
llama-4-maverickLlama 4 Maverickmeta128K$0.27$0.85general, multilingual, open-source
deepseek-v3.2DeepSeek V3.2deepseek128K$0.28$0.42general, code, multilingual, value

​
Mid

The everyday workhorses — strong quality at a sensible price across most tasks.
IDModelProviderContextInputOutputStrengths
grok-4.1Grok 4.1xai2M$0.20$0.50realtime, news, long-context
gpt-5-miniGPT-5 Miniopenai128K$0.25$2.00structured, json, tool-use
qwen-2.5-72bQwen 2.5 72Balibaba128K$0.30$0.80math, chinese, multilingual, code
kimi-k2Kimi K2moonshot128K$0.35$1.50long-context, chinese, analysis
mistral-medium-3Mistral Medium 3mistral131K$0.40$2.00multilingual, european, code
gemini-3-flashGemini 3 Flashgoogle1M$0.50$3.00multimodal, fast, long-context
claude-haiku-4.5Claude Haiku 4.5anthropic200K$1.00$5.00instruction, safety, chat, fast
sonarSonarperplexity128K$1.00$1.00realtime, news, search, citations

​
Premium

High-accuracy reasoning, writing, and code for higher-stakes work.
IDModelProviderContextInputOutputStrengths
deepseek-r1DeepSeek R1deepseek128K$0.55$2.19reasoning, math, chain-of-thought
gpt-5.2GPT-5.2openai128K$1.75$14.00reasoning, agentic, structured, code
gemini-3.1-proGemini 3.1 Progoogle1M$2.00$12.00multimodal, long-context, reasoning
claude-sonnet-4.5Claude Sonnet 4.5anthropic1M$3.00$15.00writing, code, nuance, instruction
sonar-proSonar Properplexity200K$3.00$15.00realtime, news, search, citations, deep-search

​
Frontier

Maximum capability for the hardest problems — reserved for when accuracy is paramount.
IDModelProviderContextInputOutputStrengths
grok-4.1-heavyGrok 4.1 Heavyxai2M$3.00$15.00deep-reasoning, realtime, science
claude-opus-4.6Claude Opus 4.6anthropic1M$5.00$25.00complex-reasoning, writing, safety
gpt-5.2-proGPT-5.2 Proopenai128K$21.00$168.00advanced-reasoning, science, legal

​
Using a specific model

Pass an id from the tables above instead of auto to skip routing: 
client.chat.completions.create(
    model="claude-sonnet-4.5",
    messages=[{"role": "user", "content": "Draft a launch announcement"}],
)
Reference these IDs in a BRAIN.md model, rules, blocked, or fallback field to shape routing without touching code.