NOUMENOpen app
NOUMEN

stepfun · Open weights

Step 3.7 Flash Thinking

Added May 29, 2026Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

Step 3.7 Flash Thinking is StepFun's high-efficiency multimodal MoE model with visible reasoning enabled for deeper agentic coding, long-context reasoning, tool use, and native image/video understanding. ⚠️ Note: This model routes through StepFun, so privacy and logging guarantees may be limited.

  • Reasoning
  • Tools
  • Vision
  • Structured output
  • Video input
Input
Text, Images, Video
Output
Text
Reasoning effort
none · low · medium · high
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.20
Output
$1.15
Cache read
$0.04

Reference prices as published by the model provider. API example

Specifications

Context window
262.1K
Max output
256K
Average output (7 days)
7.1K tokens
Longer than 99% of 468 models

Side by side

Step 3.7 Flash Thinking compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Step 3.7 Flash ThinkingstepfunThis modelStep 5 PreviewstepfunRemove
Providerstepfunstepfun
Discovery tierOpen weightsBalanced
AddedMay 29, 2026Sep 20, 2026
Limits
Context window262.14K1MBest
Max output256KBest65.54K
Average output7.1K tokens1.5K tokens
Pricing, USD per 1M tokens
Input$0.20Best$1.00
Output$1.15Best$2.70
Cache read$0.04Best$0.05
Performance
Request success (24h)100.0%100.0%
Benchmark indexes
Intelligence Index—43.7
Capabilities
ReasoningYesYes
ToolsYesYes
VisionYesYes
Structured outputYesYes
Video inputYesYes
Parallel toolsNoYes
Reasoning effortnone, low, medium, highlow, medium, high

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Step 3.7 Flash Thinking
  • Step 5 Preview

Intelligence Index

  • Step 3.7 Flash Thinking—
  • Step 5 Preview43.7
3.8Catalog models by score55.1

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 51.0%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 17.7%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 1,423 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 1,587 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 14.8%

  • AA-LCRv1.1

    Reasoning across long documents

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 88.3%

  • MLCR-AA

    Long-context reasoning over medical records

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 18.3%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 46.5%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 33.3%

  • SciCode

    Scientific computing in Python

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 58.9%

Legacy evaluations

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Step 3.7 Flash Thinking: —
    • Step 5 Preview: 88.3%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Prompt cache hits
0.0%
About $0.20 per 1M input tokens with caching
Request success (24h)
100.0%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 100.0%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "stepfun/step-3.7-flash:thinking",
    "reasoning_effort": "medium",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="stepfun/step-3.7-flash:thinking",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

stepfun

Other stepfun models, starting with the same family.