Advertise here
#250

Step 5 Preview

Stepfun Release: 2026-10-09 Tested on: 2026-10-09 03:52 stepfun/step-5-preview::medium
600B total (27B active)MoEClosed
(high) (medium) (low)

Summary

Step 5 Preview scores 6.1 on AI BENCHY and ranks #250. It has 9.6 reliability, a 53.6% pass rate, $4.136 total cost, and 136.10s average response time.

What makes Step 5 Preview unique: It stands out most in Instructions following, where it ranks #1, while Coding is its weakest area at #16.

Model facts

Researched on 2026-10-09

Reported
Parameters
600B total (27B active)
Architecture
MoE
Availability
Closed
License
-

OpenRouter reports 600B total / 27B active sparse MoE parameters for this exact API checkpoint. Vendor documentation confirms 1M context, 64K output and low/medium/high reasoning controls. No official downloadable exact checkpoint and license were verified in this registration; community mirrors alone do not establish an official open-source release. OpenRouter declares mandatory reasoning with medium default.

Parameter source Availability source

Score

6.1

Consistency

8.2

Reliability

9.6

Total Cost (Current Price)

$4.136

Total Output Tokens

1,434,621

Total Input Tokens

261,842

Input Price

$1.000 / 1M

Output Price

$2.700 / 1M

Cache Read Price

$0.050 / 1M

Cache Write Price

N/A

Cache prices apply to input tokens. Reads reuse cached prompts; writes store them and can cost extra. Output tokens use the output price.

Tests Correct

Wrong Tests: 13

Attempt pass rate: 53.6%

Flaky tests

5

Flaky tests had mixed outcomes across runs (at least one pass and one fail).

Response Time (avg)

136.10s

Response Time (max): 438.09s

Response Time (total): 3130.27s

No answer: 6 Wrong answer: 3 Did not follow instructions: 2 API error: 1 Invalid tool call: 1

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#250 Step 5 Preview

medium
Cost
$0.042
Time
90.3s
Tokens
15,574 tok

Charts

Choose the first model, then click a second model to open a side-by-side page.

Total Output Tokens

Score vs Total Output Tokens

Quick Compare

Step 5 PreviewmediumvsQwen3.5 Plus 2026-04-20noneStep 5 PreviewmediumvsGranite 4.2 8BmediumStep 5 PreviewmediumvsInklingnoneFree AvailableStep 5 PreviewmediumvsKAT-Coder-Pro V2.5defaultStep 5 PreviewmediumvsApodex 1.1 MinilowFree AvailableStep 5 PreviewmediumvsApodex 1.1 MinimediumFree AvailableStep 5 PreviewmediumvsTrinity Large ThinkingmediumStep 5 PreviewmediumvsGLM 5.3 FlashXlowStep 5 PreviewmediumvsGemini 3.5 FlashnoneStep 5 PreviewmediumvsMiMo-V2.6-Pronone

Category Breakdown

Category Score Consistency Tests Correct
Agentic 6.1 3.1
Anti-AI Tricks 6.4 7.9
Coding 2.8 7.2
Combined 7.3 5.8
Data parsing and extraction 10.0 10.0
Domain specific 2.9 7.2
General Intelligence 3.4 9.3
Instructions following 10.0 10.0
Puzzle Solving 7.8 10.0
Tool Calling 10.0 10.0
Trivia 3.0 10.0

Compared models