Skip to main content

Model comparison

MiMo-V2.5 vs Qwen3.5 397B

Data verified

Head-to-head evidence from 7 shared benchmark results across 3 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

Xiaomi
58.62/100
Margin
1.6pts
← winning
57.01/100
1 category wins2 category wins

Public leaderboard positions: MiMo-V2.5 #62 (Estimated); Qwen3.5 397B #71 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. MiMo-V2.5 and Qwen3.5 397B share 7 comparable benchmark results. 3 of 8 categories are comparable. 4 results are unique to MiMo-V2.5; 48 to Qwen3.5 397B.

Updated July 21, 2026
Shared results
7
MiMo-V2.5 only
4
Qwen3.5 397B only
48
Comparable categories
3 / 8

Pick MiMo-V2.5 if you want the stronger benchmark profile. Qwen3.5 397B only becomes the better choice if coding is the priority or you would rather avoid the extra latency and token burn of a reasoning model.

Confidence note. This is a partial-evidence comparison with 7 shared benchmark results across 3 evidence categories; 3 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

MiMo-V2.5 has the cleaner BenchAlign overall profile here, landing at 58.62 versus 57.01. It is a real lead, but still close enough that category-level strengths matter more than the headline number.

MiMo-V2.5's sharpest advantage is in agentic, where it averages 65.8 against 56.5. The single biggest benchmark swing on the page is Terminal-Bench 2.0, 65.8% to 52.5%. Qwen3.5 397B does hit back in coding, so the answer changes if that is the part of the workload you care about most.

MiMo-V2.5 is the reasoning model in the pair, while Qwen3.5 397B is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. MiMo-V2.5 gives you the larger context window at 1M, compared with 128K for Qwen3.5 397B.

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 MiMo-V2.5 and Qwen3.5 397B
CategoryMiMo-V2.5ΔQwen3.5 397B
CodingMiMo-V2.556.1Margin 10.4Qwen3.5 397B66.5
AgenticMiMo-V2.565.8Margin 9.3Qwen3.5 397B56.5
MultimodalMiMo-V2.579.0Margin 0.6Qwen3.5 397B79.6
ReasoningMiMo-V2.5Not measuredMarginNo overlapQwen3.5 397B63.2
KnowledgeMiMo-V2.5Not measuredMarginNo overlapQwen3.5 397B56.6
MathMiMo-V2.5Not measuredMarginNo overlapQwen3.5 397B90.6
MultilingualMiMo-V2.5Not measuredMarginNo overlapQwen3.5 397B84.7
Inst. FollowingMiMo-V2.5Not measuredMarginNo overlapQwen3.5 397B92.6

Decisive benchmark drivers

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

More
A · MiMo-V2.5B · Qwen3.5 397B
  1. Terminal-Bench 2.0

    Agentic
    Source ↗
    A 65.8%B 52.5%
    Winner: MiMo-V2.5Δ 13.3
    Terminal-Bench 2.0: MiMo-V2.5 scored 65.8%; Qwen3.5 397B scored 52.5%. MiMo-V2.5 wins this benchmark.
  2. SWE-bench Pro

    Coding
    Source ↗
    A 56.1%B 50.9%
    Winner: MiMo-V2.5Δ 5.2
    SWE-bench Pro: MiMo-V2.5 scored 56.1%; Qwen3.5 397B scored 50.9%. MiMo-V2.5 wins this benchmark.
  3. MMMU-Pro

    Multimodal
    Source ↗
    A 77.9%B 79%
    Winner: Qwen3.5 397BΔ 1.1
    MMMU-Pro: MiMo-V2.5 scored 77.9%; Qwen3.5 397B scored 79%. Qwen3.5 397B wins this benchmark.
  4. CharXiv

    Multimodal
    Source ↗
    A 81%B 80.8%
    Winner: MiMo-V2.5Δ 0.2
    CharXiv: MiMo-V2.5 scored 81%; Qwen3.5 397B scored 80.8%. MiMo-V2.5 wins this benchmark.

Operational comparison

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

MetricMiMo-V2.5Qwen3.5 397BComparison
Input / output priceUSD per 1M tokensMiMo-V2.5Not availableQwen3.5 397B$0.6 input / $3.6 outputA complete price comparison is not available.
Generation speedtokens per secondMiMo-V2.5Not availableQwen3.5 397B96 tok/sA complete speed comparison is not available.
First-answer latencyseconds to first tokenMiMo-V2.5Not availableQwen3.5 397B2.44 sA complete latency comparison is not available.
Context windowmaximum listed tokensMiMo-V2.51MQwen3.5 397B128KMiMo-V2.5 lists the larger context window.

Benchmark Deep Dive

AgenticMiMo-V2.5 wins
BenchmarkMiMo-V2.5Qwen3.5 397BResult
Claw-EvalSource 62.3%56.8%MiMo-V2.5 leads
MM-ClawBenchSource 23.8%Not comparable
Terminal-Bench 2.0Source 65.8%52.5%MiMo-V2.5 leads
Gert LabsSource 46.89%46.76%MiMo-V2.5 leads
ResearchClawBenchSource 16.9%14.2%MiMo-V2.5 leads
BrowseCompSource 62%Not comparable
QwenClawBenchSource 51.8%Not comparable
τ³-bench resultsSource 68.4%Not comparable
VITA-BenchSource 43.7%Not comparable
DeepPlanningSource 37.6%Not comparable
ToolathlonSource 36.3%Not comparable
MCP AtlasSource 46.1%Not comparable
MCP-TasksSource 74.2%Not comparable
WideResearchSource 74.0%Not comparable
τ²-bench resultsSource 95.6%Not comparable
AA Agentic IndexSource 19.9%Not comparable
APEX-Agents-AASource 15.3%Not comparable
GDPval-AASource 23.1%Not comparable
GDPval-AASource 962Not comparable
CodingQwen3.5 397B wins
BenchmarkMiMo-V2.5Qwen3.5 397BResult
SWE-bench ProSource 56.1%50.9%MiMo-V2.5 leads
Terminal-Bench 2.0Source 65.8%Not comparable
SWE-bench VerifiedSource 76.2%Not comparable
LiveCodeBench v6Source 83.6%Not comparable
AA-SciCodeSource 42.0%Not comparable
AA Coding IndexSource 48.2%Not comparable
Reasoning
BenchmarkMiMo-V2.5Qwen3.5 397BResult
LongBench v2Source 63.2%Not comparable
AI-NeedleSource 68.7%Not comparable
AA-LCRSource 65.7%Not comparable
CritPtSource 1.7%Not comparable
Knowledge
BenchmarkMiMo-V2.5Qwen3.5 397BResult
GPQASource 88.4%Not comparable
SuperGPQASource 70.4%Not comparable
MMLU-ProSource 87.8%Not comparable
MMLU-ReduxSource 94.9%Not comparable
C-EvalSource 93%Not comparable
HLESource 28.7%Not comparable
Artificial Analysis Intelligence IndexSource 33.7%Not comparable
AA-GPQA DiamondSource 89.3%Not comparable
AA-HLESource 27.3%Not comparable
AA-Omniscience IndexSource -29.8%Not comparable
AA-Omniscience AccuracySource 31.4%Not comparable
AA-Omniscience Hallucination RateSource 89.1%Not comparable
Math
BenchmarkMiMo-V2.5Qwen3.5 397BResult
AIME26Source 93.3%Not comparable
HMMT Feb 2025Source 94.8%Not comparable
HMMT Nov 2025Source 92.7%Not comparable
HMMT Feb 2026Source 87.9%Not comparable
MMAnswerBenchSource 80.9%Not comparable
Multilingual
BenchmarkMiMo-V2.5Qwen3.5 397BResult
MMLU-ProXSource 84.7%Not comparable
NOVA-63Source 59.1%Not comparable
MultimodalQwen3.5 397B wins
BenchmarkMiMo-V2.5Qwen3.5 397BResult
Video-MME (with subtitle)Source 87.7%Not comparable
CharXivSource 81%80.8%MiMo-V2.5 leads
MMMU-ProSource 77.9%79%Qwen3.5 397B leads
Design Arena WebsiteSource 1291Not comparable
MathVisionSource 88.6%Not comparable
VideoMMMUSource 84.7%Not comparable
ScreenSpot ProSource 65.6%Not comparable
V*Source 95.8%Not comparable
AA-MMMU-ProSource 77.3%Not comparable
Inst. Following
BenchmarkMiMo-V2.5Qwen3.5 397BResult
IFEvalSource 92.6%Not comparable
AA-IFBenchSource 78.8%Not comparable
Frequently Asked Questions (4)

Which is better, MiMo-V2.5 or Qwen3.5 397B?

MiMo-V2.5 is ahead on BenchLM's BenchAlign leaderboard, 58.62 to 57.01. The biggest single separator in this matchup is Terminal-Bench 2.0, where the scores are 65.8% and 52.5%.

Which is better for coding, MiMo-V2.5 or Qwen3.5 397B?

Qwen3.5 397B has the edge for coding in this comparison, averaging 66.5 versus 56.1. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, MiMo-V2.5 or Qwen3.5 397B?

MiMo-V2.5 has the edge for agentic tasks in this comparison, averaging 65.8 versus 56.5. Inside this category, Terminal-Bench 2.0 is the benchmark that creates the most daylight between them.

Which is better for multimodal and grounded tasks, MiMo-V2.5 or Qwen3.5 397B?

Qwen3.5 397B has the edge for multimodal and grounded tasks in this comparison, averaging 79.6 versus 79. Inside this category, MMMU-Pro is the benchmark that creates the most daylight between them.

Related Comparisons

Last updated: July 21, 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.