Step 5 Preview

StepFun's 600B-parameter (27B active) sparse MoE flagship for long-horizon agentic and coding work, priced at $1/M input tokens.

StepFun · Frontier/foundation models · model

Published in Feed

Overview

StepFun launched Step 5 Preview on September 20, 2026, a 600-billion-parameter sparse mixture-of-experts model with 27B parameters active per token, a 1M-token context window, and native text, image, and video input, aimed at software engineering, professional knowledge work, and finance. It scored 44 on the Artificial Analysis Intelligence Index against a similar-price-tier median of 24. In a 24-hour agent experiment, it tuned an H100 kernel to 508 TFLOPS versus 493 for Claude Opus 5, and separately raised Qwen3-30B-A3B's AIME24 score from 53.3% to 60% through automated post-training. StepFun opened paid API access the same day and has a countdown for open weights on October 15, 2026.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

  • 600B total parameters, 27B active per token (sparse MoE)
  • 1M-token context window with native text, image, and video input
  • Artificial Analysis Intelligence Index score of 44 vs. a similar-price-tier median of 24
  • Demonstrated automated kernel tuning (508 TFLOPS on H100) and automated post-training gains in agent experiments
  • Open weights planned for October 15, 2026

Use cases

  • Long-horizon agentic software engineering tasks
  • Professional knowledge work and finance workflows
  • Automated ML engineering tasks like kernel tuning and post-training

Pricing

$1.00/million input tokens (cache miss), $0.05/million (cache hit), $2.70/million output tokens (reasoning tokens billed as output) via StepFun's API.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released
20 Sept 2026
Record updated
24 Sept 2026
Record last checked
24 Sept 2026

Explore the Feed

Other entries filed under frontier/foundation models; inclusion does not imply equivalent features.

  • Muse Wonder

    Meta's most capable Muse models, with active memory and long-goal focus for agents, 1M context and $4/$20 per million tokens.

  • Rho-1

    Reka's 19B omni-reasoning model that handles text, images, video and robot actions in one network, in research preview.

  • Amazon Nova 3

    Amazon's third-generation Nova frontier model family, unveiled at an AWS keynote with a 1M-token context window and a dedicated agentic tool-use variant.

Browse the Feed