25 signals
HOlO V1 IS LIVEone ranked AI digest a day, scored in publicREAD HOW IT WORKS →HOlO V2 STARTSyour X account, your signals, every day

Model registry / StepFun / Step / Step 3.5 Flash 2603

Step 3.5 Flash 2603

StepFun flash model for efficient multimodal reasoning, coding, and tool use

Line

Step

Weights

Open

Released

2026-04-02

Context

256K

Input

$0.10 / 1M

Output

$0.30 / 1M

Max output

256K

Coverage

not covered yet

Overview

Step 3.5 Flash 2603 is a flash model from StepFun for efficient multimodal reasoning, coding, and tool use, according to the maker. It is the latest release in the step line, dated 2026-04-02, succeeding Step 3.5 Flash from 2026-01-29. It has a context window of 256,000 tokens and a maximum output of 256,000 tokens. The weights are open. It supports reasoning mode and tool calling. It accepts text and returns text. The knowledge cutoff is 2025-01. It is served by 5 hosts. Pricing starts at $0.10 per million input tokens and $0.30 per million output tokens, with cached input at $0.02 per million tokens.

Specs

Released2026-04-02
LineStepFun · Step
WeightsOpen
Context256K tokens
Max output256K tokens
Inputtext
Outputtext
ReasoningYes
Tool callingYes
Knowledge cutoff2025-01
API idstep-3.5-flash-2603

Pricing

Input$0.10 / 1M tokens
Cached input$0.02 / 1M tokens
Output$0.30 / 1M tokens

Against Step 3.5 Flash: input $0.09 to $0.10.

Strengths

  • 256k context and output window
  • Open weights for self-hosting
  • Reasoning mode and tool calling
  • Input from $0.10 per million tokens
  • Cached input from $0.02 per million tokens

Best for

  • Reach for it for coding and tool use
  • Reach for it for long-context text reasoning
  • Reach for it for self-hosted deployment with open weights
  • Reach for it for text generation with large output windows

How to access

ProviderModel id
NanoGPTstep-3.5-flash-2603
StepFun (China)step-3.5-flash-2603
StepFun (Global)step-3.5-flash-2603
StepFun Step Plan (China)step-3.5-flash-2603
StepFun Step Plan (Global)step-3.5-flash-2603

5 hosts serve this model at the price above · the maker's documentation

Step: every version

VersionReleasedContextInput
Step 5 PreviewCurrent2026-09-161M$0.959
Step 3.7 Flash2026-05-29256K$0.185
Step 3.5 Flash 26032026-04-02256K$0.10
Step 3.5 Flash2026-01-29256K$0.09

The registry keeps the six most recent versions of a line. Older ones stay in the file and can be surfaced without a rebuild.

FAQ

What is the context window and output limit?
Both the context window and the maximum output are 256,000 tokens.
Is the model open weights?
Yes, the weights are open.
What are the prices?
Input from $0.10 per million tokens, output $0.30 per million, cached input $0.02 per million.
Open weightsReasoning256K context