Step-3.7-Flash

Step 3.7 Flash

Step 3.7 Flash is an open-source frontier multimodal reasoning model by StepFun. Built on a sparse Mixture of Experts (MoE) architecture, it activates only ~11B of its 198B total parameters per token, and pairs its language backbone with a vision encoder for native image understanding — delivering state-of-the-art reasoning at a fraction of the cost of dense models.

Capabilities

• Multimodal: Native image understanding — send text and images together via the standard image_url content format
• Reasoning: Extended thinking with **\<think>** blocks, with selectable depth via reasoning_effort (low, medium, high). Reasoning is always on for this model
• Tool Calling: Native function calling support with parallel tool invocation
• Long Context: 256K token context window
• Structured Output: JSON via response_format

Benchmarks

Step 3.7 Flash delivers strong results across search-heavy benchmarks. It scores 47.20% on HLE with Tools, up from 35.68% (text-only) for Step 3.5 Flash, and outperforms Flash models from DeepSeek V4 and Gemini 3.5. It reaches 75.82% on BrowseComp, approaching larger models such as Claude Opus 4.7 and GLM 5.1. On DeepSearchQA, it achieves 92.82% F1 score, comparable to Kimi K2.6, a 1T / 32B-active model. On ResearchRubrics, it scores 71.68%, ahead of GPT 5.5 at 61.50% and close to Claude Opus 4.7 at 73.92%. These results show that Step 3.7 Flash combines Flash-level efficiency with strong deep-retrieval and research capabilities.

Architecture

Usage

from openai import OpenAI

client = OpenAI(
 base_url="https://api.deepinfra.com/v1/openai",
 api_key="YOUR_DEEPINFRA_TOKEN",
)

# Chat with reasoning
response = client.chat.completions.create(
 model="stepfun-ai/Step-3.7-Flash",
 messages=[{"role": "user", "content": "Prove that sqrt(2) is irrational."}],
)
print(response.choices[0].message.reasoning_content) # thinking
print(response.choices[0].message.content) # answer

# Image understanding (multimodal)
response = client.chat.completions.create(
 model="stepfun-ai/Step-3.7-Flash",
 messages=[{"role": "user", "content": [
 {"type": "text", "text": "What is in this image?"},
 {"type": "image_url", "image_url": {"url": "https://example.com/photo.jpg"}},
 ]}],
)

# Control reasoning depth
response = client.chat.completions.create(
 model="stepfun-ai/Step-3.7-Flash",
 messages=[{"role": "user", "content": "Plan a 3-day trip to Tokyo."}],
 extra_body={"reasoning_effort": "high"},
)

# Tool calling
response = client.chat.completions.create(
 model="stepfun-ai/Step-3.7-Flash",
 messages=[{"role": "user", "content": "What's the weather in Paris?"}],
 tools=[{
 "type": "function",
 "function": {
 "name": "get_weather",
 "description": "Get weather for a city",
 "parameters": {
 "type": "object",
 "properties": {"city": {"type": "string"}},
 "required": ["city"],
 },
 },
 }],
)

Links

- https://huggingface.co/stepfun-ai/Step-3.7-Flash
- https://github.com/stepfun-ai/Step-3.7-Flash
copy