Model Comparison

MAI-Thinking-1 vs Step-3.5-FlashWhich is better in 2026?

Step-3.5-Flash shows notably better performance in the majority of benchmarks.

Verdict: MAI-Thinking-1 vs Step-3.5-Flash — which is better?

MAI-Thinking-1 (by Microsoft) and Step-3.5-Flash (by StepFun) are two of the AI models people compare most. Here is how they stack up on benchmarks, price and capabilities, and which one to pick in 2026.

MAI-Thinking-1 outperforms in 1 benchmarks (LiveCodeBench v6), while Step-3.5-Flash is better at 3 benchmarks (AIME 2025, SWE-Bench Verified, Terminal-Bench 2.0). Step-3.5-Flash shows notably better performance in the majority of benchmarks.

Choose MAI-Thinking-1 if…

  • you want the most recent training data — it shipped Jun 2026

Choose Step-3.5-Flash if…

  • you want the strongest raw capability — it leads on 3 of 4 shared benchmarks
  • you need open weights you can self-host or fine-tune

Performance Benchmarks

Comparative analysis across standard metrics

4 benchmarks

MAI-Thinking-1 outperforms in 1 benchmarks (LiveCodeBench v6), while Step-3.5-Flash is better at 3 benchmarks (AIME 2025, SWE-Bench Verified, Terminal-Bench 2.0).

Step-3.5-Flash shows notably better performance in the majority of benchmarks.

Tue Jul 21 2026 • llm-stats.com

Arena Performance

Human preference votes

Model Size

Parameter count comparison

804.0B diff

MAI-Thinking-1 has 804.0B more parameters than Step-3.5-Flash, making it 410.2% larger.

Microsoft
MAI-Thinking-1
1.0Tparameters
StepFun
Step-3.5-Flash
196.0Bparameters
1000.0B
MAI-Thinking-1
196.0B
Step-3.5-Flash

Context Window

Maximum input and output token capacity

Only Step-3.5-Flash specifies input context (65,536 tokens). Only Step-3.5-Flash specifies output context (8,192 tokens).

Microsoft
MAI-Thinking-1
Input- tokens
Output- tokens
StepFun
Step-3.5-Flash
Input65,536 tokens
Output8,192 tokens
Tue Jul 21 2026 • llm-stats.com

License

Usage and distribution terms

MAI-Thinking-1 is licensed under a proprietary license, while Step-3.5-Flash uses Apache 2.0.

License differences may affect how you can use these models in commercial or open-source projects.

MAI-Thinking-1

Proprietary

Closed source

Step-3.5-Flash

Apache 2.0

Open weights

Release Timeline

When each model was launched

MAI-Thinking-1 was released on 2026-06-02, while Step-3.5-Flash was released on 2026-02-02.

MAI-Thinking-1 is 4 months newer than Step-3.5-Flash.

MAI-Thinking-1

Jun 2, 2026

1 months ago

4mo newer
Step-3.5-Flash

Feb 2, 2026

5 months ago

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

Microsoft
Higher LiveCodeBench v6 score (87.7% vs 86.4%)
StepFun
Larger context window (65,536 tokens)
Has open weights
Higher AIME 2025 score (97.3% vs 97.0%)
Higher SWE-Bench Verified score (74.4% vs 73.5%)
Higher Terminal-Bench 2.0 score (51.0% vs 46.0%)

Detailed Comparison

Interactive Arena

Judge for yourself.

Run your own prompts against MAI-Thinking-1 and Step-3.5-Flash side-by-side, then vote on the output you prefer.

MAI-Thinking-1
✓ Preferred
Step-3.5-Flash
Open in Playground
AI Model Comparison Table
Feature
Microsoft
MAI-Thinking-1
StepFun
Step-3.5-Flash

FAQ

Common questions about MAI-Thinking-1 vs Step-3.5-Flash.

Which is better, MAI-Thinking-1 or Step-3.5-Flash?

Step-3.5-Flash shows notably better performance in the majority of benchmarks. MAI-Thinking-1 is made by Microsoft and Step-3.5-Flash is made by StepFun. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.

How does MAI-Thinking-1 compare to Step-3.5-Flash in benchmarks?

MAI-Thinking-1 scores LongFact: 98.0%, AIME 2025: 97.0%, AIME 2026: 94.5%, GraphWalks: 90.0%, AIR-Bench: 88.0%. Step-3.5-Flash scores AIME 2025: 97.3%, Tau-bench: 88.2%, LiveCodeBench v6: 86.4%, IMO-AnswerBench: 85.4%, SWE-Bench Verified: 74.4%.

What are the context window sizes for MAI-Thinking-1 and Step-3.5-Flash?

MAI-Thinking-1 supports an unknown number of tokens and Step-3.5-Flash supports 66K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between MAI-Thinking-1 and Step-3.5-Flash?

Key differences include licensing (Proprietary vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes MAI-Thinking-1 and Step-3.5-Flash?

MAI-Thinking-1 is developed by Microsoft and Step-3.5-Flash is developed by StepFun.

Related comparisons