Step 5 Preview
Step 5 PreviewAlternatives?
CheaperClaude Haiku 5.5max×3.8-79%$0.21Remove sampling parameters+5
BetterMuse Glimmer 30BhighOpen+20%-94%$0.058Context: 1M → 131K
OpenGLM 5.3 Flashmax×3.1-72%$0.28Defaults to max · Self-hosted+4
CheaperGPT-6.1 Solmax×2-29%$0.72Remove unsupported parameters+5CheaperMuse Spark 1.2xhigh+87%-5%$0.97Only auto tool_choice is supported+1Claude Haiku 5.5max×3.8-79%$0.21Remove sampling parameters+5GLM 5.3 FlashmaxOpen×3.1-72%$0.28Defaults to max · Self-hosted+4GPT-6.1 Solmax×2-29%$0.72Remove unsupported parameters+5Muse Spark 1.2xhigh+87%-5%$0.97Only auto tool_choice is supported+1GPT-5.6 Terraxhigh+30%-38%$0.63Prompts with >272K input tokensMuse Glimmer 30BhighOpen+20%-94%$0.058Context: 1M → 131KGPT-5.6 Lunamax+17%-83%$0.18Claude Haiku 5.5max×3.8-79%$0.21Remove sampling parameters+5GLM 5.3 FlashmaxOpen×3.1-72%$0.28Defaults to max · Self-hosted+4GPT-5.6 Terraxhigh+30%-38%$0.63Prompts with >272K input tokensGLM 5.3 FlashmaxOpen×3.1-72%$0.28Defaults to max · Self-hosted+4GLM 5.3maxOpen×2.9+95%$2.01Memory (4-bit): 363 GB → 452 GB+5
Qwen3.8 27BxhighOpen+30%same$1.01Context: 1M → 262K+5Muse Glimmer 30BhighOpen+20%-94%$0.058Context: 1M → 131K
MiMo v2.6 ProthinkingOpen+10%-87%$0.13Memory (4-bit): 363 GB → 615 GB+4Run it yourself
estimateBenchmarks
in Chat & KnowledgeWatch out 3
web_search not available · Chat Completions›
Built-in Chat Completions web_search is unavailable; integrate external search through a function tool and return the results. Source ↗
Output includes reasoning›
Output billing includes the reasoning process as well as the final answer. Source ↗
LRU eviction during peak traffic›
LRU eviction removes unused cache entries more aggressively during peak traffic, so cached prefixes may not survive. Source ↗
FAQ
How much does Step 5 Preview cost?›
$1 per 1M input tokens and $2.70 per 1M output tokens; cached input $0.10 per 1M; cheapest via stepfun at $0.96 / $2.74. A typical benchmark task costs about $1.03 at default effort.
Is Step 5 Preview free to use?›
There is no free API; the cheapest paid price is $0.96 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Step 5 Preview, and what is it best at?›
#20 of 334 in Chat & Knowledge, #18 of 132 in Coding, #34 of 133 in Agents & Tool Use, #65 of 195 in Writing & Language, #40 of 124 in Vision & Documents. Strongest at: Coding, Agents & work tasks, Reasoning, Long context, Knowledge.
Is Step 5 Preview better than MiniMax M3?›
On MLCR-AA, Step 5 Preview scores 16.7 vs 17.2 — about the same. It is 102% more expensive per task.
How can I use Step 5 Preview?›
Through the API of DeepInfra, Featherless, InferenceNet, Novita, StepFun.
What is the context window of Step 5 Preview?›
1M tokens, with up to 64K output tokens.
When was Step 5 Preview released? Is there a newer version?›
Released 2026-09-16. It is the current version in its line.
Is Step 5 Preview open source?›
Yes, the weights are open.
Can I run Step 5 Preview locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Step 5 Preview read images?›
It accepts text and images, video as input.
What is Step 5 Preview?›
Step 5 Preview is an open-weights language model from StepFun, released Sep 2026.
What are the best alternatives to Step 5 Preview?›
Claude Haiku 5.5 (cheaper), Muse Glimmer 30B (better), GLM 5.3 Flash (open weights).