Skip to main content

Model comparison

Qwen3.7 Plus vs Step 3.7 Flash

Data verified

Head-to-head evidence from 23 shared benchmark results across 6 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

67.22/100
Margin
16.4pts
← winning
50.87/100
2 category wins0 category wins

Public leaderboard positions: Qwen3.7 Plus #21 (Supported); Step 3.7 Flash #110 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Qwen3.7 Plus and Step 3.7 Flash share 23 comparable benchmark results. 2 of 8 categories are comparable. 46 results are unique to Qwen3.7 Plus; 6 to Step 3.7 Flash.

Updated July 23, 2026
Shared results
23
Qwen3.7 Plus only
46
Step 3.7 Flash only
6
Comparable categories
2 / 8

Pick Qwen3.7 Plus if you want the stronger benchmark profile. Step 3.7 Flash only becomes the better choice if its workflow or ecosystem matters more than the raw scoreboard.

Confidence note. This is a partial-evidence comparison with 23 shared benchmark results across 6 evidence categories; 2 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

Qwen3.7 Plus is clearly ahead on the BenchAlign aggregate, 67.22 to 50.87. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

Qwen3.7 Plus's sharpest advantage is in coding, where it averages 75.6 against 56.3. The single biggest benchmark swing on the page is Terminal-Bench 2.0, 70.3% to 59.5%.

Qwen3.7 Plus gives you the larger context window at 1M, compared with 256K for Step 3.7 Flash.

Category breakdown

Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.

Category scores and score margins for Qwen3.7 Plus and Step 3.7 Flash
CategoryQwen3.7 PlusΔStep 3.7 Flash
CodingQwen3.7 Plus75.6Margin 19.3Step 3.7 Flash56.3
AgenticQwen3.7 Plus71.7Margin 5.3Step 3.7 Flash66.4
ReasoningQwen3.7 Plus91.7MarginNo overlapStep 3.7 FlashNot measured
KnowledgeQwen3.7 Plus60.1MarginNo overlapStep 3.7 FlashNot measured
MathQwen3.7 Plus92.9MarginNo overlapStep 3.7 FlashNot measured
MultilingualQwen3.7 Plus85.4MarginNo overlapStep 3.7 FlashNot measured
MultimodalQwen3.7 Plus81.5MarginNo overlapStep 3.7 FlashNot measured
Inst. FollowingQwen3.7 Plus84.5MarginNo overlapStep 3.7 FlashNot measured

Decisive benchmark drivers

The largest measured benchmark gaps in this matchup, with exact reported values.

More
A · Qwen3.7 PlusB · Step 3.7 Flash
  1. Terminal-Bench 2.0

    Agentic
    Source ↗
    A 70.3%B 59.5%
    Winner: Qwen3.7 PlusΔ 10.8
    Terminal-Bench 2.0: Qwen3.7 Plus scored 70.3%; Step 3.7 Flash scored 59.5%. Qwen3.7 Plus wins this benchmark.
  2. SWE-bench Pro

    Coding
    Source ↗
    A 57.6%B 56.3%
    Winner: Qwen3.7 PlusΔ 1.3
    SWE-bench Pro: Qwen3.7 Plus scored 57.6%; Step 3.7 Flash scored 56.3%. Qwen3.7 Plus wins this benchmark.

Operational comparison

Runtime and commercial metrics are compared only when both models have a complete sourced value.

MetricQwen3.7 PlusStep 3.7 FlashComparison
Input / output priceUSD per 1M tokensQwen3.7 PlusNot availableStep 3.7 Flash$0.2 input / $1.15 outputA complete price comparison is not available.
Generation speedtokens per secondQwen3.7 PlusNot availableStep 3.7 FlashNot availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenQwen3.7 PlusNot availableStep 3.7 FlashNot availableA complete latency comparison is not available.
Context windowmaximum listed tokensQwen3.7 Plus1MStep 3.7 Flash256KQwen3.7 Plus lists the larger context window.

Benchmark Deep Dive

AgenticQwen3.7 Plus wins
BenchmarkQwen3.7 PlusStep 3.7 FlashResult
Terminal-Bench 2.0Source 70.3%59.5%Qwen3.7 Plus leads
QwenClawBenchSource 61.8%Not comparable
QwenWebBenchSource 1536Not comparable
Claw-EvalSource 62.7%67.1%Step 3.7 Flash leads
BFCL v4Source 72.9%Not comparable
MCP AtlasSource 73.2%Not comparable
VITA-BenchSource 45.6%Not comparable
DeepPlanningSource 62.3%Not comparable
OSWorld-VerifiedSource 73.3%Not comparable
AndroidWorldSource 81.0%Not comparable
AA Agentic IndexSource 20.8%21.5%Step 3.7 Flash leads
APEX-Agents-AASource 22.4%14.8%Qwen3.7 Plus leads
τ²-bench resultsSource 93%98.5%Step 3.7 Flash leads
GDPval-AASource 21.8%25.9%Step 3.7 Flash leads
GDPval-AASource 9361017Step 3.7 Flash leads
OSWorld 2.0Source 2.8%Not comparable
BrowseCompSource 75.8%Not comparable
DeepSearchQASource 92.8%Not comparable
ToolathlonSource 49.5%Not comparable
HLE w/ toolsSource 47.2%Not comparable
Gert LabsSource 51.57%Not comparable
CodingQwen3.7 Plus wins
BenchmarkQwen3.7 PlusStep 3.7 FlashResult
Terminal-Bench 2.0Source 70.3%59.5%Qwen3.7 Plus leads
SWE-bench VerifiedSource 77.7%Not comparable
SWE-bench ProSource 57.6%56.3%Qwen3.7 Plus leads
SWE MultilingualSource 75.8%Not comparable
NL2RepoSource 41.1%Not comparable
SciCodeSource 51.3%Not comparable
LiveCodeBenchSource 89.6%Not comparable
AA Coding IndexSource 55.9%39.6%Qwen3.7 Plus leads
AA-SciCodeSource 45.5%40.0%Qwen3.7 Plus leads
Reasoning
BenchmarkQwen3.7 PlusStep 3.7 FlashResult
CritPtSource 9.1%2.3%Qwen3.7 Plus leads
MRCRv2Source 91.7%Not comparable
AA-LCRSource 65.0%63.7%Qwen3.7 Plus leads
Knowledge
BenchmarkQwen3.7 PlusStep 3.7 FlashResult
GPQASource 90.3%Not comparable
GPQA-DSource 90.3%Not comparable
HLESource 34.7%Not comparable
MMLU-ProSource 88.5%Not comparable
MMLU-ReduxSource 94.5%Not comparable
SuperGPQASource 71.4%Not comparable
MMMLUSource 89.0%Not comparable
Artificial Analysis Intelligence IndexSource 39.0%30.3%Qwen3.7 Plus leads
AA-GPQA DiamondSource 90.0%80.9%Qwen3.7 Plus leads
AA-HLESource 33.4%19.9%Qwen3.7 Plus leads
AA-Omniscience IndexSource 2.4%-37.5%Qwen3.7 Plus leads
AA-Omniscience AccuracySource 22.2%25.4%Step 3.7 Flash leads
AA-Omniscience Hallucination RateSource 25.5%84.4%Qwen3.7 Plus leads
Math
BenchmarkQwen3.7 PlusStep 3.7 FlashResult
HMMT Feb 2026Source 92.9%Not comparable
IMOAnswerBenchSource 86.0%Not comparable
ApexSource 22.7%Not comparable
Multilingual
BenchmarkQwen3.7 PlusStep 3.7 FlashResult
MMLU-ProXSource 85.4%Not comparable
NOVA-63Source 58.8%Not comparable
INCLUDESource 83.0%Not comparable
MAXIFESource 88.8%Not comparable
PolyMathSource 84.0%Not comparable
Multimodal
BenchmarkQwen3.7 PlusStep 3.7 FlashResult
MMMU-ProSource 79%Not comparable
MathVisionSource 90.3%Not comparable
CharXivSource 85.9%Not comparable
ERQASource 69.8%Not comparable
MedXpertQA (MM)Source 71.0%Not comparable
ScreenSpot ProSource 79.0%Not comparable
SimpleVQASource 81.7%79.2%Qwen3.7 Plus leads
MMSearch-PlusSource 41.4%Not comparable
RealWorldQASource 86.9%Not comparable
OmniDocBench 1.5Source 91.4%Not comparable
OCRBench V2Source 70.7%Not comparable
ODINW13Source 51.1%Not comparable
Video-MME (with subtitle)Source 88.0%Not comparable
VideoMMMUSource 85.4%Not comparable
MLVU (M-Avg)Source 87.4%Not comparable
AA-MMMU-ProSource 80.5%75.3%Qwen3.7 Plus leads
Design Arena WebsiteSource 12881211Qwen3.7 Plus leads
V*Source 95.3%Not comparable
Inst. Following
BenchmarkQwen3.7 PlusStep 3.7 FlashResult
IFEvalSource 94.6%Not comparable
IFBenchSource 79.1%Not comparable
AA-IFBenchSource 78.0%67.3%Qwen3.7 Plus leads
Frequently Asked Questions (3)

Which is better, Qwen3.7 Plus or Step 3.7 Flash?

Qwen3.7 Plus is ahead on BenchLM's BenchAlign leaderboard, 67.22 to 50.87. The biggest single separator in this matchup is Terminal-Bench 2.0, where the scores are 70.3% and 59.5%.

Which is better for coding, Qwen3.7 Plus or Step 3.7 Flash?

Qwen3.7 Plus has the edge for coding in this comparison, averaging 75.6 versus 56.3. Inside this category, AA Coding Index is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, Qwen3.7 Plus or Step 3.7 Flash?

Qwen3.7 Plus has the edge for agentic tasks in this comparison, averaging 71.7 versus 66.4. Inside this category, GDPval-AA is the benchmark that creates the most daylight between them.

Related Comparisons

Last updated: July 23, 2026

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.