Skip to main content
The short answer: tell the agent your use case and priority, let it pick. The long answer, so you can sanity-check what the agent suggests:

By use case

Use caseRecommended starting pointWhy
Customer-support chatbotLlama 3.1 8B with a compatible 4-bit variant on L40SLow latency, solid instruction following
Document summarizationQwen 2.5 14B with a compatible 4-bit or FP8 variantLong context (128K), strong at compression
Code generationDeepSeek Coder V2 on a high-throughput GPUStrong open coder, use FP8 only where the compatibility check passes
Multilingual chatQwen 2.5 7B on L40SNative multilingual training; outperforms Llama on non-English
Reasoning / mathDeepSeek R1 on H100Best open reasoning model
Fast extraction / classificationPhi-3 Mini on L4Cheapest, still good enough for JSON extraction
Voice assistantWhisper Large V3 Turbo + Llama 8B + XTTS v23-node pipeline, sub-500ms end-to-end
RAG backendbge-m3 embeddings + Llama 8Bbge-m3 for retrieval, any instruction model for generation

By priority

Pick the smallest model that still passes your quality bar. Let the optimizer try compatible 4-bit or FP8 variants. Deploy on the lowest GPU tier that clears your latency target. Enable speculation on a paid Core plan.Target: P99 under 200 ms for 1 to 8B, under 400 ms for 14 to 30B.

By model size

SizeTypical costQuality ceilingBest for
1 to 3BCheapestSimple extraction, classificationInternal tools, FAQ bots
7 to 8BLowGood chat, basic tool useProduction chat, customer support
14BMediumStrong general-purposeMost SaaS features
30 to 32BHighGreat reasoning and codeAgent backbones, code copilots
70B+HighestSOTA open performanceFlagship products, research

When to ask the agent to recommend

If you don’t know, just ask:
I need a chatbot for an e-commerce site. Budget $200/month, under 200ms latency,
traffic is ~50 RPM. What model do you recommend?
The agent will suggest, explain its reasoning, and let you adjust. See Best practices for how to phrase these asks.

Next steps

Models catalog

Full list of supported models.

GPUs and pricing

Which GPUs match which model sizes.

Optimization

Quantization, FP8, TensorRT-LLM, and GPU selection.

Example prompts

Copy-ready prompts for every shape.