Step 3.5 Flash
Step 3.5 Flash is behind the leaders in coding, instruction following, reasoning, long context, agentic tasks, and factuality. Too few results yet to rate safety, math, multimodal tasks, or multilingual tasks.
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 Safety, Math, Multimodal or Multilingual.
Price
$0.10input$0.30outputper million tokens
From Artificial Analysis · 2 providers tracked · All prices
Evidence
58results on36benchmarks
- 6 independently verified
- 29 aggregator
- 23 vendor-reported
From 9 sources · latest Oct 7, 2026 · How verification works
API features
Tool callingReasoning
As listed by OpenRouter
Step 3.5 Flash benchmark results
58 results on 36 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.
28.4% behind the leader4 of 10 ranked benchmarks measured
- 86.40Oct 6, 2026
- 74.40Oct 6, 2026
- SciCode40.37Sep 4, 2026aa_scicode
- Terminal-Bench Hard27.27Oct 7, 2026aa_terminalbench_hard
Show 2 more coding resultsHide 2 coding results
- SciCode38.54Sep 4, 2026aa_scicode
- Terminal-Bench Hard32.58Oct 7, 2026aa_terminalbench_hard
32.0% behind the leader1 of 3 ranked benchmarks measured
- IFBench64.56Oct 7, 2026aa_ifbench
Show 1 more instruction following resultHide 1 instruction following result
- IFBench66.53Oct 7, 2026aa_ifbench
32.1% behind the leader3 of 6 ranked benchmarks measured
- GPQA Diamond83.13Oct 7, 2026gpqa
- Humanity's Last Exam21.13Oct 7, 2026aa_hle
- 2.47Oct 7, 2026
Show 5 more reasoning resultsHide 5 reasoning results
- 2.29Oct 7, 2026
- GPQA Diamond82.63Oct 7, 2026gpqa
- 83.50May 3, 2026
- Humanity's Last Exam24.47Oct 7, 2026aa_hle
- 23.10May 3, 2026
36.6% behind the leader1 of 3 ranked benchmarks measured
- 50.00Oct 7, 2026
Show 1 more long context resultHide 1 long context result
- 63.00Oct 7, 2026
37.1% behind the leader2 of 7 ranked benchmarks measured
- 69.00Oct 6, 2026
- 27.60Jun 15, 2026
Show 1 more agentic resultHide 1 agentic result
- 51.60May 30, 2026
37.2% behind the leader2 of 4 ranked benchmarks measured
- AA-Omniscience · Accuracy23.62Oct 7, 2026omniscienceAccuracy
- AA-Omniscience · Non-hallucination14.34Oct 7, 2026omniscienceNonHallucination
Show 2 more factuality resultsHide 2 factuality results
- AA-Omniscience · Accuracy24.62Oct 7, 2026omniscienceAccuracy
- AA-Omniscience · Non-hallucination8.69Oct 7, 2026omniscienceNonHallucination
0 of 5 ranked benchmarks measured
- 44.64Sep 2, 2026
- 86.36Sep 2, 2026
- 96.67Sep 2, 2026
Show 5 more math resultsHide 5 math results
- 96.67May 2, 2026
- AIME 202696.67May 3, 2026MathArena AIME 2026
- 86.36May 10, 2026
- HMMT Feb 202686.36May 10, 2026MathArena HMMT Feb 2026
- 85.40Oct 6, 2026
More results
Benchmarks outside the capability baskets. They are not ranked against a leader.
- 94.17May 10, 2026
- -41.82Oct 7, 2026
- AA Intelligence16.62Oct 7, 2026aa_intelligence_index
- τ²-Bench Telecom (AA run)94.44Oct 7, 2026aa_tau2
- τ²-Bench Telecom (AA run)87.43Oct 7, 2026aa_tau2
- AA Intelligence16.96Oct 7, 2026aa_intelligence_index
Show 19 more resultsHide 19 results
- 52.00Jun 18, 2026
- 48.23Jun 18, 2026
- -44.22Oct 7, 2026
- 97.30Oct 6, 2026
- 31.64Jun 18, 2026
- Artificial Analysis Coding Index34.56Jun 18, 2026aa_coding_index
- browsecomp_with_context_manager69.00May 30, 2026BrowseComp (w/ Context Manager)
- 66.90May 30, 2026
- 73.70May 30, 2026
- 84.50May 30, 2026
- HMMT Feb. 202598.40May 30, 2026HMMT 2025 (Feb.)
- HMMT Nov. 202594.00May 30, 2026HMMT 2025 (Nov.)
- 86.36May 3, 2026
- 84.40May 3, 2026
- 65.30May 30, 2026
- 51.00Oct 6, 2026
- xbench-DeepSearch56.30May 30, 2026xbench-DeepSearch (2025.10)
- 56.30May 3, 2026
- 88.20Jun 12, 2026
Step 3.5 Flash: common questions
Who makes Step 3.5 Flash?
Step 3.5 Flash is made by StepFun.
When was Step 3.5 Flash released?
Step 3.5 Flash was released on Feb 2, 2026, according to Artificial Analysis.
What is Step 3.5 Flash good at?
Step 3.5 Flash is behind the leaders in coding, instruction following, reasoning, long context, agentic tasks, and factuality. Too few results yet to rate safety, math, multimodal tasks, or multilingual tasks.
How much does Step 3.5 Flash cost?
Step 3.5 Flash costs $0.10 per million input tokens and $0.30 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 is cheaper than 84% of the 328 priced models we track.
How many benchmarks has Step 3.5 Flash been tested on?
We track 58 results for Step 3.5 Flash on 36 benchmarks from 9 sources, 6 of them independently verified. The latest was recorded on Oct 7, 2026.
Which API features does Step 3.5 Flash support?
OpenRouter lists tool calling and reasoning for Step 3.5 Flash.
About this record
Where Step 3.5 Flash's numbers come from, and every name it appears under.
- Tracked since
- Apr 27, 2026
- Newest source mention
- Aug 23, 2026
Where the results come from
Verification: 58 scores · 6 independently verified · 29 aggregator-attributed · 23 vendor-reported. How these tiers are assigned
From 9 sources on 4 sites. Artificial Analysis supplies 29 of them; the 6 independently verified results come from 1 site. Bars are coloured by trust tier.
- artificialanalysis.ai29
- huggingface.co17
- api.llm-stats.com6
- matharena.ai6