Step 5 Preview
Step 5 Preview is strong in long context, capable in factuality and reasoning, and behind the leaders in multimodal tasks and agentic tasks. Too few results yet to rate coding, safety, math, multilingual tasks, or instruction following.
Capability profile
Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.
Too few results yet to rate Coding, Safety, Math, Multilingual or Instruction Following.
Price
$1.00input$2.70outputper million tokens
From Artificial Analysis · 2 providers tracked · All prices
Evidence
16results on15benchmarks
- 3 independently verified
- 13 aggregator
From 4 sources · latest Oct 16, 2026 · How verification works
API features
Tool callingStructured outputsReasoning
As listed by OpenRouter
Step 5 Preview benchmark results
16 results on 15 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.
4.0% behind the leader1 of 3 ranked benchmarks measured
- 88.33Oct 16, 2026
19.8% behind the leader2 of 4 ranked benchmarks measured
- AA-Omniscience · Non-hallucination56.98Oct 16, 2026omniscienceNonHallucination
- AA-Omniscience · Accuracy41.53Oct 16, 2026omniscienceAccuracy
20.3% behind the leader2 of 6 ranked benchmarks measured
- Humanity's Last Exam46.48Oct 16, 2026aa_hle
- 20.86Oct 16, 2026
27.4% behind the leader1 of 6 ranked benchmarks measured
- MMMU-Pro76.36Oct 16, 2026aa_mmmu_pro
Show 1 more multimodal resultHide 1 multimodal result
- 1264.58Oct 3, 2026
29.4% behind the leader3 of 7 ranked benchmarks measured
- 54.33Oct 16, 2026
- Terminal-Bench 4.033.33Oct 16, 2026
Show 2 more agentic resultsHide 2 agentic results
- 37.98Oct 16, 2026
- 55.56Oct 16, 2026
0 of 10 ranked benchmarks measured
- 1571.04Sep 26, 2026
- SciCode58.91Oct 16, 2026aa_scicode
More results
Benchmarks outside the capability baskets. They are not ranked against a leader.
- AA Intelligence29.00Sep 19, 2026Artificial Analysis Intelligence Index
- AA Intelligence43.73Oct 16, 2026aa_intelligence_index
- 16.38Oct 16, 2026
Step 5 Preview: common questions
Who makes Step 5 Preview?
Step 5 Preview is made by StepFun.
When was Step 5 Preview released?
Step 5 Preview was released on Sep 18, 2026, according to Artificial Analysis.
What is Step 5 Preview good at?
Step 5 Preview is strong in long context, capable in factuality and reasoning, and behind the leaders in multimodal tasks and agentic tasks. Too few results yet to rate coding, safety, math, multilingual tasks, or instruction following.
How much does Step 5 Preview cost?
Step 5 Preview costs $1.00 per million input tokens and $2.70 per million output tokens, according to Artificial Analysis. We track its price at 2 providers. At a mix of three input tokens to one output token, it costs more than 65% of the 323 priced models we track.
How many benchmarks has Step 5 Preview been tested on?
We track 16 results for Step 5 Preview on 15 benchmarks from 4 sources, 3 of them independently verified. The latest was recorded on Oct 16, 2026.
Which API features does Step 5 Preview support?
OpenRouter lists tool calling, structured outputs, and reasoning for Step 5 Preview.
About this record
Where Step 5 Preview's numbers come from, and every name it appears under.
- Tracked since
- Sep 19, 2026
- Newest source mention
- Oct 17, 2026
Where the results come from
Verification: 16 scores · 3 independently verified · 13 aggregator-attributed. How these tiers are assigned
From 4 sources on 2 sites. Artificial Analysis supplies 14 of them; the 3 independently verified results come from 2 sites. Bars are coloured by trust tier.
- artificialanalysis.ai14
- datasets-server.huggingface.co2