Evidence checked Nov 11, 2026

Verified Nov 11, 2026 via Step 3.5 Flash: step-3-5-flash-projectAI models

Step 3.5 Flash

Step 3.5 Flash is an open-source foundation model aimed at reasoning and agentic tasks, built as a sparse Mixture of Experts that activates 11B of its 196B parameters per token.

Verified facts
7
Offers
1
Sources
1
Pricing
Free offer

Overview

What Step 3.5 Flash covers

Step 3.5 Flash is an open-source foundation model from StepFun built on a sparse Mixture of Experts architecture that activates 11B of its 196B parameters per token. The README describes it as aimed at reasoning, coding and agentic tasks.

It states a 256K context window using a 3:1 sliding window attention ratio, 3-way multi-token prediction, and local deployment on hardware such as the Mac Studio M4 Max and NVIDIA DGX Spark. It lists vLLM, SGLang, Hugging Face Transformers and llama.cpp as supported backends.

Read the full editorial overview

Licence and running it

The repository licence file is the Apache License, Version 2.0, and the README's license section says the project is open-sourced under Apache 2.0.

The README also describes cloud API access through supported providers, which have their own terms, and notes that the China platform requires +86 phone number verification. It does not state prices.

Maintenance and what is not established

The repository has no GitHub releases, so this capture follows the default branch. Its latest commit is dated 3 April 2026 and the repository is not archived.

The README's known-issues section says the model relies on longer generation trajectories than Gemini 3.0 Pro for comparable quality and may be less stable in specialized domains and long multi-turn dialogues. Nothing was installed or tested.

Sources

Step 3.5 Flash: step-3-5-flash-project · Evidence: documentation · Not checked

Plans & prices

Prices as listed by Step 3.5 Flash, before any tax or fee the pricing page does not state.

Scroll the table sideways for every column.

PlanRate period not statedCoverage limits
Plan: Self-hosted
FreeTax: Confirm at checkoutUpfront: not established
ConditionsDated publisher project README captured on 9 November 2026 (commit 21d85a5f); supplier description, not a deployment or independent test. The license line covers the project repository. The README also lists OpenRouter free and standard tiers and StepFun platform endpoints for API access; those prices were not captured. Hardware, compute, accounts and hosted APIs have separate costs and terms.
Unknown fees: Hardware and compute for running the modelVerified Nov 19, 2026
The llama.cpp guide lists int4 GGUF weights of 111.5 GB, about 7 GB of runtime overhead and a minimum of 120 GB VRAM, with 128 GB of unified memory recommended.
ConditionsDated publisher project README captured on 9 November 2026 (commit 21d85a5f); supplier description, not a deployment or independent test. System requirements of the llama.cpp route only; other backends and quantizations were not checked.
Verified Nov 11, 2026

Step 3.5 Flash: step-3-5-flash-project

Documented facts

Each fact comes from official documentation and shows when it was verified.

  • Who it is for

    Step 3.5 Flash is an open-source foundation model aimed at reasoning and agentic tasks, built as a sparse Mixture of Experts that activates 11B of its 196B parameters per token.

    Verified Nov 8, 2026 · Step 3.5 Flash: step-3-5-flash-project

  • Coverage limits

    The README says Step 3.5 Flash is tailored for coding and work-centric tasks and may be less stable in specialized domains or long multi-turn dialogues, with repetitive reasoning or mixed-language outputs.

    Verified Nov 8, 2026 · Step 3.5 Flash: step-3-5-flash-project

  • API model identifier

    step-3.5-flash

    Dated publisher project README captured on 9 November 2026 (commit 21d85a5f); supplier description, not a deployment or independent test. API model ID used in the README's agent-platform guides; OpenRouter uses stepfun/step-3.5-flash.

    Verified Nov 11, 2026 · Step 3.5 Flash: step-3-5-flash-project

  • Context window

    256K tokens in the context window

    Dated publisher project README captured on 9 November 2026 (commit 21d85a5f); supplier description, not a deployment or independent test. Supplier statement of the supported context window; no long-context test was run.

    Verified Nov 11, 2026 · Step 3.5 Flash: step-3-5-flash-project

  • Availability

    Local inference is documented with vLLM, SGLang, Hugging Face Transformers and llama.cpp.

    Dated publisher project README captured on 9 November 2026 (commit 21d85a5f); supplier description, not a deployment or independent test. Documentation of local backends, not a performance claim or a hosted-service entitlement; cloud API access through OpenRouter and the StepFun platform is described separately.

    Verified Nov 11, 2026 · Step 3.5 Flash: step-3-5-flash-project

Applies to the facts above: Dated publisher project README captured on 7 November 2026; supplier description, not a deployment or independent test. Infrastructure, accounts, third-party services and managed offerings have separate costs and terms.

Not confirmed in our sources

We have not confirmed these terms for Step 3.5 Flash in the linked documentation. That does not mean Step 3.5 Flash lacks them; confirm them before you rely on them.

  • Provider
  • Modalities
  • Maximum output
  • Input token rate
  • Output token rate

Sources & methodology

Facts come from the linked documentation. Documentation checks are not hands-on tests, and readable sources can have incomplete coverage.

Retrieved Nov 11, 2026

Discovered an undocumented discrepancy or pricing shift? Submit a correction request.

Submit correction

Step 3.5 Flash alternatives

Other AI models tools to weigh. Different billing units do not establish a cost winner, so compare plans side by side.

Step 3.5 Flash comparisons

Compare two products in the same category using their documented terms and sources.