LLM Pricing
ModelsLeaderboardsToolsProvidersReleasesArticles

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list

Step 5 Preview

stepfun·stepfun/step-5-preview·GA·Open weights·NEW
StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window

Specs & pricing

Input / output per 1M tokens
Official price·StepFun (China)
$1.05 / $2.99
Blended $1.53 · Cache read $0.10
Lowest paid·EzkielynaGateway
$0.46 / $1.24
Blended $0.65 · 2.5× spread
3 more $0 channels
Context
1,000,000
Output limit
65,536
Knowledge cutoff
—
Released / updated
2026-09-16 / 2026-09-20
Capabilities
✓ Reasoning✓ Tool use✓ Structured output✓ Temperature✓ Attachments
⚠ Providers report capability flags inconsistently
Modalities
TextImageVideoAudio
Weights
Hugging Face

Step 5 Preview by stepfun is offered by 26 providers on this page. Public prices are shown for 21 of them. Its official list price is $1.05 per 1M input tokens and $2.99 per 1M output tokens. The lowest paid channel is Ezkielyna at $0.458 / $1.24 per 1M, about 2.3× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it. It also has 3 free ($0) channels and 2 covered by a paid subscription; free tiers usually carry rate limits, and subscription-covered access bills $0 per token only after the subscription fee.

Artificial Analysis rates it 28.5 on the Intelligence Index, with 44.2 for Strategy & Ops, its strongest domain. Against its lowest blended price of $0.652 per 1M, that is roughly 44 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 166 tokens per second, with 0.76s to the first token. Running one task of the Artificial Analysis suite costs about $2.3453, which reflects how many tokens its reasoning consumes rather than the unit price alone.

The context window is 1,000,000 tokens at the reference host, but hosts report different limits, from 1,000,000 to 1,048,576, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. Accepted input modalities are Text, Image, Video, and Audio. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use.

Quality & performanceArtificial Analysis · Intelligence Index v5

Intelligence28.5
Value44
Engineering37
Finance & Accounting40.5
Legal38.6
Healthcare & Medical39.9
Strategy & Ops44.2
Economics38.6
Output speed166 tok/s
TTFT0.76 s
Cost per task$2.35
Value formulaIQ 28.5 ÷ min blended $0.652 = 44

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 26 providers21 with public prices · 3 free · 2 subscription-covered

  • ProviderInput / Output
  • NvidiaCloud
    Free
    Context 1,048,576 ⚠Output limit 1,048,576
  • OpenCodeGateway
    step-5-preview-free
    Free
    Cache read $0Context 1,000,000Output limit 65,536deprecated
  • OpenCode GoGateway
    step-5-preview-free
    Free
    Cache read $0Context 1,000,000Output limit 65,536deprecated
  • EzkielynaGateway
    $0.46 / $1.24
    Cache read $0.046Context 1,000,000Output limit 1,000,000
  • RunInfraGateway
    stepfun-ai/Step-5-Preview
    $0.90 / $2.43
    Cache read $0.09Context 1,048,576 ⚠Output limit 32,768
  • StepFun (Global)OfficialFirst-party
    $1.00 / $2.70
    Cache read $0.10Context 1,024,000 ⚠Output limit 65,536
  • Kilo GatewayGateway
    $1.00 / $2.70
    Cache read $0.10Context 1,000,000Output limit 64,000
  • DevPass (LLM Gateway)Gateway
    $1.00 / $2.70
    Cache read $0.10Context 1,000,000Output limit 1,000,000
  • Deep InfraCloud
    stepfun-ai/Step-5-Preview
    $1.00 / $2.70
    Cache read $0.10Context 1,048,576 ⚠Output limit 1,000,000deprecated
  • OpenRouterGateway
    $1.00 / $2.70
    Cache read $0.10Context 1,000,000Output limit 64,000
ProviderTierInputOutputCache readCache writeContextOutput limitStatus
NvidiaCloudFree——1,048,576 ⚠1,048,576
OpenCode
step-5-preview-free
GatewayFree$0—1,000,00065,536deprecated
OpenCode Go
step-5-preview-free
GatewayFree$0—1,000,00065,536deprecated
EzkielynaGateway$0.46$1.24$0.046—1,000,0001,000,000
RunInfra
stepfun-ai/Step-5-Preview
Gateway$0.90$2.43$0.09—1,048,576 ⚠32,768
StepFun (Global)OfficialFirst-party$1.00$2.70$0.10—1,024,000 ⚠65,536
Kilo GatewayGateway$1.00$2.70$0.10—1,000,00064,000
DevPass (LLM Gateway)Gateway$1.00$2.70$0.10—1,000,0001,000,000
Deep Infra
stepfun-ai/Step-5-Preview
Cloud$1.00$2.70$0.10—1,048,576 ⚠1,000,000deprecated
OpenRouterGateway$1.00$2.70$0.10—1,000,00064,000

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Reasoning control

effort = noneeffort = minimaleffort = loweffort = mediumeffort = higheffort = xhigheffort = max

Interleaved thinking (reasoning between tool calls) is declared by 10 of 26 providers.

2 / 26 providers expose no reasoning control (reasoning_options: []).

Your usage cost

Common scenarios
1Ezkielyna$103.93
2RunInfra$204.30
3Requesty$221.00
4NanoGPT$221.00
5Eden AI$221.00
6CommandCode$221.00
7StepFun (Global) · Official$227.00

3 more channels offer $0 (Nvidia, OpenCode, OpenCode Go); free tiers usually have rate limits and no SLA, excluded from ranking.

Switch to Ezkielyna to save $123.07/mo (54%)
Note: this is a gateway; verify availability and rate limits yourself.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$2.99$02026-09-202026-11-15
2026-11-15Input list $1.05Output list $2.99Min blended $0.65

Artificial Analysis evaluations26 items

Humanity's Last Exam46.5%
HLE-Diamond24.3%
MMMU-Pro76.4%
CritPt20.9%
CritPt v2 Hard32.5%
SciCode58.9%
FrontierCode38.4%
FrontierCode 2.012.0%
Terminal-Bench 4.033.3%
Terminal-Bench-Science2.4%
AutomationBench-AA51.0%
APEX-Agents-AA38.0%
AA-AnalystAgent35.0%
EnterpriseOps-Gym-AA47.2%
ITBench-AA55.6%
AA-Briefcase1423
GDPval-AA1592
GDP.pdf14.8%
AA-LCR88.3%
MLCR-AA18.3%
AA-Omniscience Index16.4
AA-Omniscience Accuracy36.4%
AA-Omniscience Non-Hallucination68.7%
AA-3D859
Code2Video-AA821
mka0.2758

Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and scale (percent correct, points or an index), so rows are not comparable with one another.

Related models

GLM-5.3-Flashcheaper alternative$0.15 / $0.50DeepSeek V4.1 Flash 1017cheaper alternative$0.15 / $0.60GPT-5.6 Lunacheaper alternative$0.20 / $1.20MiniMax-M3cheaper alternative$0.30 / $1.20Step 3.7 Flashsame lab$0.20 / $1.21Step 3.5 Flashsame lab$0.10 / $0.31
Data partly from models.dev (MIT) · StepFun (China) official docs ↗