Model details
Step 3.5 Flash 2603
Step 3.5 Flash 2603 is a text-only model from StepFun's "Flash" family, designed to handle long-context reasoning and agentic workflows at a low price point. Source documentation positions it as supporting advanced reasoning for complex multi-step tasks and as capable of calling external tools and functions, making it well suited to chain-of-thought workloads, function-calling agents, and multi-turn pipelines that need to ingest large documents or codebases in a single pass. Its 256,000-token context window and matching 256,000-token maximum output allow it to keep very long conversations, retrieval-augmented inputs, or whole-file analyses in memory without aggressive truncation, which is its most distinctive practical advantage over smaller-context peers in the same Flash tier.
In practice, the model fits workflows where throughput and cost matter more than frontier reasoning quality: high-volume agentic services, batch processing of long documents, and tool-using assistants that repeatedly hit APIs. Independent benchmarking lists Step 3.5 Flash 2603 as one provider offering with measurable output speed, and the same source marks it as deprecated on that benchmark site, with StepFun's newer Step 3.7 Flash recommended for projects that want actively maintained performance numbers. For teams that need the open-weights path the catalog advertises, Step 3.5 Flash 2603 remains a budget-friendly option for the use cases above, while teams prioritizing long-term support may prefer migrating to the Step 3.7 Flash successor.
Quick Info
Powered by- Provider
- StepFun (Global)
- Model key
- step-3.5-flash-2603
- Release date
- Apr 2, 2026
- Last updated
- Apr 2, 2026
- Knowledge cutoff
- 2025-01
- Input modalities
- Output modalities
- Capabilities
Cost
- Input token cost
- $0.10
- Output token cost
- $0.30
Limits
- Input tokens
- 256,000 tokens
- Output tokens
- 256,000 tokens
- Context window
- 256,000 tokens