StepFunReleased May 28, 2026
Open weights
step-3.7-flash
Its best results are in Knowledge Work (#85 of 176) and Mathematical Reasoning (#90 of 176); weakest in Reliability (#115 of 177). Three of seven dimensions are estimated, because no results are published yet.
Composite IQ ?
103
Effective cost
$1.33/1M
$0.20 in · $1.15 out
Speed
112tok/s
25s typical response
Context
262Ktokens
262,000 tokens in
Evidence
4of 7 dims
10 of 39 benchmarks
IQ profile
Seven dimensions, averaged equally into Composite IQ
MeasuredEstimated95% range
#1 in dimensionTop-10 median
IQ dimensions
Ranks are among models scored in each dimension. Ranges are model-based estimates, not calibrated confidence intervals.
DimensionIQPosition · 70–160RankEvidencevs Top 10
step-3.7-flash95% range (model-based, uncalibrated)Top-10 median#1 in dimension100 = average human
Benchmark results
10 of 39 Composite IQ benchmarks have a published result for step-3.7-flash. Benchmarks without a result stay listed so gaps are visible.
Abstract Reasoning IQ 85 · estimatedNo published results0/5
No published results yet for Abstract Reasoning. The 85 shown is estimated from step-3.7-flash's other dimensions and will be replaced as results come in. How estimates work
ARC-AGI-3ARC-AGI-2ARC-AGI-1Chess PuzzlesMystery Game Puzzles
Mathematical Reasoning IQ 111 · measuredBest: #44 of 64 on MathArena1/5
BenchmarkScorePercentileRankLeader · source
MathArena
Full source leaderboard reconciliation · Oct 28
18.7%
#44 / 64
Leader gpt-6.1-sol · 84.0%
Full source leaderboard reconciliation · Oct 28
Full source leaderboard reconciliation · Oct 28
Not yet reported
FrontierMath Tier 4FrontierMath Tier 1-3ProofBench 1.1AIME
Academic Reasoning IQ 119 · measuredBest: #96 of 167 on SciCode5/6
BenchmarkScorePercentileRankLeader · source
SciCode
Artificial Analysis v4.3.2 · Oct 27
43.9%
#96 / 167
Leader opus-5.5 · 66.9%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
MMMU-Pro
Artificial Analysis v4.3.2 · Oct 27
75.3%
#52 / 89
Leader fable-5.1 · 90.6%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
CritPt
Artificial Analysis v4.3.2 · Oct 27
2.3%
#102 / 159
Leader Muse Wonder Full · 32.9%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
GPQA Diamond
Artificial Analysis v4.3.2 · Oct 27
80.9%
#98 / 147
Leader gpt-6-astra · 96.1%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
Humanity's Last Exam
Artificial Analysis v4.3.2 · Oct 27
21.4%
#116 / 167
Leader opus-5.5 · 61.4%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
Not yet reported
MMLU-Pro
Programmatic Reasoning IQ 92 · estimatedNo published results0/6
No published results yet for Programmatic Reasoning. The 92 shown is estimated from step-3.7-flash's other dimensions and will be replaced as results come in. How estimates work
LiveCodeBenchIOI V2Terminal-Bench 4.0ProgramBenchSWE-rebenchMirrorCode
Computer Use IQ 99 · estimatedNo published results0/5
No published results yet for Computer Use. The 99 shown is estimated from step-3.7-flash's other dimensions and will be replaced as results come in. How estimates work
BrowseCompOSWorld-VerifiedMCP AtlasArena.ai Agent ArenaAgents' Last Exam
Reliability IQ 109 · measuredBest: #56 of 96 on IFBench3/6
BenchmarkScorePercentileRankLeader · source
IFBench
Artificial Analysis v4.3.2 · Oct 27
67.3%
#56 / 96
Leader minimax-m3 · 82.9%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
AA Long Context Reasoning v1.1
Artificial Analysis v4.3.2 · Oct 27
73.7%
#95 / 154
Leader kimi-k3 · 88.7%
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
AA Omniscience
Artificial Analysis v4.3.2 · Oct 27
-37.3
#134 / 158
Leader opus-5.5 · 46.4
Artificial Analysis v4.3.2 · Oct 27
Artificial Analysis v4.3.2 · Oct 27
Not yet reported
SimpleQA VerifiedBullshitBench v2FACTS Grounding
Knowledge Work IQ 110 · measuredBest: #84 of 124 on GDPval-AA v2.1 normalized score1/6
BenchmarkScorePercentileRankLeader · source
GDPval-AA v2.1 normalized score
Artificial Analysis v4.3.2 · Oct 20
17.8%
#84 / 124
Leader opus-5.5 · 68.3%
Artificial Analysis v4.3.2 · Oct 20
Artificial Analysis v4.3.2 · Oct 20
Not yet reported
AA-Briefcase v1.1AutomationBench-AAGDP.pdfAA-AnalystAgentEnterpriseOps-Gym-AA
Release & availability
Release details on record for this model.
- Released
- Developer
- StepFun
- Weights
- Open weightsDownloadable and self-hostable
- Context
- 262,000 tokens
- API price
- $0.20 in · $1.15 outper 1M tokens · full pricing
About this score
How much of step-3.7-flash's IQ is measured, and how the rest is estimated.
10/ 39benchmark results
4/ 7dimensions measured
3estimated
0/5Abstract
1/5Math
5/6Academic
0/6Coding
0/5Computer Use
3/6Reliability
1/6Knowledge Work
Bar height = share of benchmarks with resultsEstimatedNo score
Measured results are shown as published
Every score in the benchmark tables comes from its source unchanged. Composite IQ averages all seven dimensions equally.
Estimates fill the gaps
Abstract Reasoning, Programmatic Reasoning and Computer Use have no published results yet, so their IQ is estimated from step-3.7-flash's other dimensions.
A missing result is not a zero
Benchmark versions and evaluation harnesses can differ between sources, so small gaps between models may not be meaningful.
Compare step-3.7-flash
Side-by-side pages for the models it is most often weighed against.
#1 overall
gpt-6.1-sol
IQ141 38 higherEffective cost$6.30 4.7× pricier
China (Other) #1
Ling 3.1 Flash
IQ120 17 higher
Closest IQ
A.X-K2
IQ103 same IQ
Most similar profile
minimax-m2.7
IQ105 2 higherEffective cost$1.46 about the same
Similar price
GLM-5.3-Flash
IQ119 16 higherEffective cost$1.33 about the same
#2 overall
opus-5.5
IQ141 38 higherEffective cost$35.73 26.8× pricier
step-3.7-flashComparison model
Frequently asked questions
Quick answers about step-3.7-flash, drawn from the same data as the rest of this page.
How AI IQ scores models →Get this data from the API →- What is the IQ of step-3.7-flash?
- step-3.7-flash has an estimated IQ of 103, which ranks #113 out of 176 public models on the AI IQ leaderboard. The score combines seven equally weighted reasoning dimensions using measured benchmark results and conservative estimates for missing coverage.
- What is step-3.7-flash best at?
- Among the dimensions with published benchmark results, step-3.7-flash ranks highest in Knowledge Work (IQ 110, #85 of 176) and Mathematical Reasoning (IQ 111, #90 of 176). Its lowest-ranked measured dimension is Reliability (IQ 109, #115 of 177).
- What is the EQ of step-3.7-flash?
- step-3.7-flash doesn't have an EQ score yet. EQ (emotional reasoning) is calculated from EQ-Bench 4 and AttuneBench, and neither has published a result for step-3.7-flash. AI IQ doesn't estimate EQ from other benchmarks, so the score appears once one of those results is published.
- Which benchmarks has step-3.7-flash been tested on?
- step-3.7-flash has published results on 10 of the 39 benchmarks behind AI IQ's Composite IQ, including SciCode, IFBench, MMMU-Pro, AA Long Context Reasoning v1.1 and CritPt.
- How much does step-3.7-flash cost?
- step-3.7-flash costs $0.20 per million input tokens and $1.15 per million output tokens. Adjusted for how many tokens it actually consumes on real tasks, its effective cost works out to $1.33 per million input-plus-output tokens.
- Is step-3.7-flash open source?
- Yes — step-3.7-flash is an open-weights model: its trained weights are published, so it can be downloaded and self-hosted.
- Who developed step-3.7-flash?
- step-3.7-flash was developed by StepFun, based in China.
- When was step-3.7-flash released?
- step-3.7-flash was publicly released on May 28, 2026.
- How large is the context window of step-3.7-flash?
- step-3.7-flash supports a context window of 262,000 tokens — the amount of text it can consider in a single request.
- How fast is step-3.7-flash?
- In independent measurements, step-3.7-flash generates about 112 tokens per second, with a typical response time of about 25 seconds.