Step 5 Preview – Benchmarks, Pricing & Intelligence Analysis
Step 5 Preview is StepFun's flagship model for agentic work, built on a sparse Mixture-of-Experts architecture (27B active / 600B total parameters). It performs strongly in software engineering and professional...
Not available on LLMBase
Step 5 Preview is not currently available in LLMBase Chat or the LLMBase Inference API. Benchmarks, pricing, and model details remain available for research and comparison.
Input Price
$1.00/1M tokens
Output Price
$2.70/1M tokens
Intelligence
43.7
Coding
N/A
Specifications
Technical details and pricing.
EU-hosted inference API
Power your AI projects with open-source models.
Drop-in OpenAI-compatible API. No data leaves Europe.
MiniMax
MiniMax M3
$0.40 / $1.40
per M tokens
Z.ai
GLM 5.3 Flash
$0.20 / $0.60
per M tokens
MoonshotAI
Kimi K3
$4.00 / $18.00
per M tokens
DeepSeek
DeepSeek V4.1 Flash
$0.40 / $1.40
per M tokens
Frequently Asked Questions
What is Step 5 Preview good for?
Use Step 5 Preview for everyday tasks like writing, summarizing, brainstorming, and getting clear explanations.
How much does Step 5 Preview cost?
Pricing is based on usage. Current rates are $1.00/1M tokens for input and $2.70/1M tokens for output.
Is Step 5 Preview available in LLMBase Chat yet?
Not yet on this page. We will add chat and inference actions automatically once an exact catalog entry exists for Step 5 Preview.
Does Step 5 Preview support images or audio?
Step 5 Preview focuses on text-based tasks.
Similar Models
Other models you might want to explore.
Step 3.7 Flash
Stepfun
Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...
Details
Claude Haiku 5.5
Anthropic
Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and...
Details
Pareto 26.10 Preview
Unbiased
Pareto is a multimodal composite model built for research, coding, and agentic workflows, while delivering frontier-level performance across a broad range of general-purpose tasks. This is a preview of the...
DetailsBenchmarks and pricing use Artificial Analysis where available. Catalog specs are used as a fallback.