Skip to main content
BenchLM
Data

Kimi K2.5 vs Qwen3.5-35B-A3B

Updated November 19, 2026. Rank says Kimi K2.5 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.

Share or export
Share on XLinkedInAPI/MCP

Decision reading

Kimi K2.5 has the higher public point estimate, 43.9 versus 34.7. Their conditional score ranges overlap. These ranges do not establish rank confidence. 19 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.

Model A
Moonshot AI logo

Moonshot AI

43.9/100

Supported · Public rank #123

90% interval 38.8–49.0

Model B
Alibaba logo

Alibaba

34.7/100

Supported · Public rank #182

Conditional range 27.6–41.7

API price per 1M tokens

Kimi K2.5: $0.60 input / $3 output

Qwen3.5-35B-A3B: No comparable API price

Context window

Kimi K2.5: 256K

Qwen3.5-35B-A3B: 262K

Released

Kimi K2.5: January 27, 2026

Qwen3.5-35B-A3B: March 4, 2026

Shared results
19
Kimi K2.5 only
49
Qwen3.5-35B-A3B only
2
Like-for-like categories
1 / 8
Supported: Kimi K2.5 and Qwen3.5-35B-A3B. Conditional ranges do not establish rank confidence.How the comparison works

Which one for your work

Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.

  • Agentic work

    Tool use, computer use, and multi-step task completion

    Kimi K2.5

    Kimi K2.5 has the higher public agentic point estimate, 39.8 to 30.8, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

    Confidence: limited
  • Long documents

    Prompts that approach the documented context limit

    Qwen3.5-35B-A3B

    Qwen3.5-35B-A3B has the larger documented context window.

    Confidence: documented
Show secondary and unsupported calls
  • Coding work

    Code generation, repair, and software-engineering tasks

    Not enough matched evidence

    Qwen3.5-35B-A3B is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.

    Confidence: limited
  • Chat turn cost

    1K fresh input + 500 output tokens

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    Confidence: listed-rates
  • Cache-heavy agent loop cost

    200K cached + 20K fresh input + 10K output tokens

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    Confidence: rate-fallback
  • Repository review cost

    50K fresh input + 3K output tokens

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    Confidence: listed-rates

Which one for a specific job

Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.

The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.

42.6Kimi K2.529.7Qwen3.5-35B-A3B

Directional only · BenchAlign v6.0

Kimi K2.5 has the higher coding point estimate. Conditional score ranges do not establish rank confidence.

Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.

Same basis rules as the category table below

Or describe your own job

The LLM Selector reads what you need a model for and returns a shortlist, ranked in the same order the job pages publish.

Describe the task and constraints. This selector recommends models; it cannot carry out the task or access your files and accounts.

We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy

What is actually comparable

Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.

5 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.

Shape of the matched evidence

Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.

Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.

Bars run 0–100 on each benchmark’s normalized display scale

Category results, on a stated basis

Each row shows the public-lane category score for both models: the BenchAlign v6.0 lane in every category, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence. Directional and not-comparable rows remain visible, but they never receive a winner.

Agentic

Like-for-like
Kimi K2.5
39.8
Supported · #133/177
Qwen3.5-35B-A3B
30.8
Supported · #157/177
Basis
BenchAlign v6.0 lane · 16 vs 4 public rows
Reading
Kimi K2.5 leads · intervals overlap

Coding

Directional only
Kimi K2.5
42.6
Supported · #116/177
Qwen3.5-35B-A3B
29.7
Estimated · #164/177
Basis
BenchAlign v6.0 lane · 10 vs 4 public rows
Reading
Directional only

Reasoning

Directional only
Kimi K2.5
44.4
Supported · #102/160
Qwen3.5-35B-A3B
35.4
Estimated · #146/160
Basis
BenchAlign v6.0 lane · 1 vs 1 public rows
Reading
Directional only

Multimodal

Directional only
Kimi K2.5
44.7
Estimated · #83/115
Qwen3.5-35B-A3B
41.2
Estimated · #96/115
Basis
BenchAlign v6.0 lane · 20 vs 7 public rows
Reading
Directional only

Knowledge

Directional only
Kimi K2.5
46.0
Estimated · #96/156
Qwen3.5-35B-A3B
36.4
Estimated · #135/156
Basis
BenchAlign v6.0 lane · 8 vs 3 public rows
Reading
Directional only

Multilingual

Directional only
Kimi K2.5
52.1
Estimated · #40/86
Qwen3.5-35B-A3B
37.1
Estimated · #71/86
Basis
BenchAlign v6.0 lane · 5 vs 1 public rows
Reading
Directional only

Instruction following

Not comparable
Kimi K2.5
42.2
Estimated · #103/142
Qwen3.5-35B-A3B
Not ranked
Basis
BenchAlign v6.0 lane · 1 vs 1 public rows
Reading
Not comparable

Math

Not comparable
Kimi K2.5
42.9
Estimated · #79/124
Qwen3.5-35B-A3B
Not ranked
Basis
BenchAlign v6.0 lane · 7 vs 0 public rows
Reading
Not comparable

Ranks count the models scored in each category’s BenchAlign v6.0 lane, so denominators differ by category. Like-for-like rows within 0.5 points read as a practical tie.

Supported evidence per lane · bars run 0–100Methodology

What each workload costs

Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.

Chat turn

1K fresh input + 500 output tokens

Kimi K2.5
$0.0021
Fits in one request
Qwen3.5-35B-A3B
Self-hosted; infrastructure cost varies
Fits in one request

Qwen3.5-35B-A3B has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

Kimi K2.5
$0.039
Fits in one request
Qwen3.5-35B-A3B
Self-hosted; infrastructure cost varies
Fits in one request

Qwen3.5-35B-A3B has no comparable published API token rate.

Cache-heavy agent loop

200K cached + 20K fresh input + 10K output tokens

Kimi K2.5
$0.162
Fits in one request
Cached input priced at the published list-input rate
Qwen3.5-35B-A3B
Self-hosted; infrastructure cost varies
Fits in one request
Cached-input rate unavailable

Kimi K2.5 has no published cached-input rate, so cached tokens use its listed input rate. Qwen3.5-35B-A3B has no comparable published API token rate.

Cached input falls back to the list input rate only where a cached rate is unpublished

Price against score, among all priced models

Overall score against output price per 1M tokens, one model per family. Pricing as of November 18, 2026.

Kimi K2.5

20406080100$0.10$0.50$1$5$10$50$100Output price, $ per 1M tokens (log)Overall score
  • 18 models score at least as high as Kimi K2.5 for the same price or less. The nearest is Gemini 3 Flash (score 53.1, also $3.00 per 1M output tokens).
  • Qwen3.5-35B-A3B has no public API price, so it is not plotted.

Gray dots are the other priced models. The dark step line is the efficiency frontier: for a model under it, another model costs no more and scores at least as high.

See Kimi K2.5 on the full price vs performance chart

Specification differences

Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.

Context window

Maximum documented context; output-token limits may be lower.

Kimi K2.5

256K

Qwen3.5-35B-A3B

262K

API model ID

Kimi K2.5

Not sourced

Qwen3.5-35B-A3B

Not sourced

Cached-input rate

A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.

Kimi K2.5

Not published

Qwen3.5-35B-A3B

No comparable hosted API rate

Provider availability

Kimi K2.5

Not sourced

Qwen3.5-35B-A3B

Not sourced

Reasoning profile

Kimi K2.5

Non-Reasoning

Qwen3.5-35B-A3B

Reasoning

Weight access

Kimi K2.5

Open Weight

Qwen3.5-35B-A3B

Open Weight

License

Kimi K2.5

Open Weight

Qwen3.5-35B-A3B

Open Weight

Release date

Kimi K2.5

2026-01-27

Qwen3.5-35B-A3B

2026-03-04

If you already use one of these models

Deployment change
The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
Quality signal
Kimi K2.5 has the higher public point estimate, 43.9 versus 34.7. Their conditional score ranges overlap. These ranges do not establish rank confidence.
Workload cost
A complete comparable API-rate estimate is not available for both models.
Context tradeoff
Qwen3.5-35B-A3B has the larger documented window (262K).
Run the same representative tasks against both endpoints before changing production traffic.

Questions

Which is better, Kimi K2.5 or Qwen3.5-35B-A3B?

Kimi K2.5 has the higher public point estimate, 43.9 versus 34.7. Their conditional score ranges overlap. These ranges do not establish rank confidence. The page therefore keeps the decision tied to the specific documented workload.

Which is better for coding, Kimi K2.5 or Qwen3.5-35B-A3B?

Kimi K2.5 scores higher for coding on the public lane, 42.6 to 29.7. Qwen3.5-35B-A3B is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; read the shared benchmark rows directly and test the models on the same task set.

Which is better for agentic tasks, Kimi K2.5 or Qwen3.5-35B-A3B?

Kimi K2.5 has the higher public agentic tasks point estimate, 39.8 to 30.8, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

Which costs less, Kimi K2.5 or Qwen3.5-35B-A3B?

Both models do not have comparable published API token rates, so this page does not name a universal price winner.

Which has the larger context window, Kimi K2.5 or Qwen3.5-35B-A3B?

Qwen3.5-35B-A3B has the larger documented context window: 262K, compared with 256K.

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

Kimi K2.5
API / mo$2,700
Self-host / mo$27,962
Break-evenNot reached at 70% load
Qwen3.5-35B-A3B
API / mo (DeepInfra)$855
Self-host / mo$298
Break-even17M/day
Model the full break-even

Benchmark evidence

The full public result ledger is available for audit without forcing a wide desktop table onto a phone.

Browse raw public benchmark evidence70 rows

Agentic

  • Terminal-Bench 2.0

    Kimi K2.550.8%
    Source
    Qwen3.5-35B-A3B40.5%
    Source

    Kimi K2.5 leads this result

  • BrowseComp

    Kimi K2.560.6%
    Source
    Qwen3.5-35B-A3B61%
    Source

    Qwen3.5-35B-A3B leads this result

  • OSWorld-Verified

    Kimi K2.563.3%
    Source
    Qwen3.5-35B-A3B54.5%
    Source

    Kimi K2.5 leads this result

  • Claw-Eval

    Kimi K2.552.3%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • QwenClawBench

    Kimi K2.554.3%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • τ³-bench results

    Kimi K2.565.7%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • DeepSearchQA

    Kimi K2.577.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • DeepPlanning

    Kimi K2.514.4%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • Toolathlon

    Kimi K2.527.8%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MCP Atlas

    Kimi K2.529.5%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MCP-Tasks

    Kimi K2.559.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • WideResearch

    Kimi K2.572.7%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • BFCL v4

    Kimi K2.568.3%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • Kimi K2.545.88%
    Qwen3.5-35B-A3B28.96%

    Kimi K2.5 leads this result

  • ResearchClawBench

    Kimi K2.514.0%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • JobBench

    Kimi K2.58.7%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

Coding

  • SWE-bench Verified

    Kimi K2.576.8%
    Source
    Qwen3.5-35B-A3B69.2%
    Source

    Kimi K2.5 leads this result

  • SWE-bench Verified*

    Kimi K2.570.8%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • LiveCodeBench v6

    Kimi K2.585.0%
    Source
    Qwen3.5-35B-A3B74.6%
    Source

    Kimi K2.5 leads this result

  • SWE-bench Pro

    Kimi K2.550.7%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • SWE Multilingual

    Kimi K2.573%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • SWE-Rebench

    Kimi K2.558.5%
    Source
    Qwen3.5-35B-A3B53.7%
    Source

    Kimi K2.5 leads this result

  • React Native Evals

    Kimi K2.577.2%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • NL2Repo

    Kimi K2.532%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • SciCode

    Kimi K2.548.7%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • Terminal-Bench 2.0

    Kimi K2.550.8%
    Source
    Qwen3.5-35B-A3B40.5%
    Source

    Kimi K2.5 leads this result

Reasoning

  • LongBench v2

    Kimi K2.561%
    Source
    Qwen3.5-35B-A3B59%
    Source

    Kimi K2.5 leads this result

Multimodal

  • MMMU

    Kimi K2.584.3%
    Source
    Qwen3.5-35B-A3B81.4%
    Source

    Kimi K2.5 leads this result

  • MMMU-Pro

    Kimi K2.578.5%
    Source
    Qwen3.5-35B-A3B75.1%
    Source

    Kimi K2.5 leads this result

  • RealWorldQA

    Kimi K2.581.0%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • OmniDocBench 1.5

    Kimi K2.588.8%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • Video-MME

    Kimi K2.587.4%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • Video-MME (with subtitle)

    Kimi K2.587.4%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • Video-MME (w/o subtitle)

    Kimi K2.583.2%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MathVision

    Kimi K2.584.2%
    Source
    Qwen3.5-35B-A3B83.9%
    Source

    Kimi K2.5 leads this result

  • We-Math

    Kimi K2.584.7%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • DynaMath

    Kimi K2.584.4%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • SimpleVQA

    Kimi K2.571.2%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MMLongBench-Doc

    Kimi K2.558.5%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • CC-OCR

    Kimi K2.579.7%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • AI2D_TEST

    Kimi K2.590.8%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • CountBench

    Kimi K2.594.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • RefCOCO (avg)

    Kimi K2.587.8%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MMVU

    Kimi K2.580.4%
    Source
    Qwen3.5-35B-A3B72.3%
    Source

    Kimi K2.5 leads this result

  • VideoMMMU

    Kimi K2.586.6%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MLVU (M-Avg)

    Kimi K2.585.0%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • CharXiv

    Kimi K2.577.5%
    Source
    Qwen3.5-35B-A3B77.5%
    Source

    Tie

  • ExtractBench (LlamaIndex)

    Kimi K2.5—
    Qwen3.5-35B-A3B88.00%
    Source

    Not directly comparable

  • V*

    Kimi K2.5—
    Qwen3.5-35B-A3B92.7%
    Source

    Not directly comparable

Knowledge

  • GPQA

    Kimi K2.587.6%
    Source
    Qwen3.5-35B-A3B84.2%
    Source

    Kimi K2.5 leads this result

  • GPQA-D

    Kimi K2.587.6%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • SuperGPQA

    Kimi K2.569.2%
    Source
    Qwen3.5-35B-A3B63.4%
    Source

    Kimi K2.5 leads this result

  • MMLU-Pro

    Kimi K2.587.1%
    Source
    Qwen3.5-35B-A3B85.3%
    Source

    Kimi K2.5 leads this result

  • MMLU-Pro (Arcee)

    Kimi K2.587.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MMLU-Redux

    Kimi K2.594.5%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • C-Eval

    Kimi K2.594%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • HLE

    Kimi K2.530.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

Multilingual

  • MMLU-ProX

    Kimi K2.582.3%
    Source
    Qwen3.5-35B-A3B81%
    Source

    Kimi K2.5 leads this result

  • NOVA-63

    Kimi K2.556.0%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • INCLUDE

    Kimi K2.583.3%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • PolyMath

    Kimi K2.543.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MAXIFE

    Kimi K2.572.8%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

Instruction following

  • IFEval

    Kimi K2.593.9%
    Source
    Qwen3.5-35B-A3B91.9%
    Source

    Kimi K2.5 leads this result

Math

  • AIME 2025

    Kimi K2.596.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • AIME26

    Kimi K2.595.8%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • AIME25 (Arcee)

    Kimi K2.596.3%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • HMMT Feb 2025

    Kimi K2.595.4%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • HMMT Nov 2025

    Kimi K2.591.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • HMMT Feb 2026

    Kimi K2.587.1%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

  • MMAnswerBench

    Kimi K2.581.8%
    Source
    Qwen3.5-35B-A3B—

    Not directly comparable

70 public results · 19 shared

Watch Kimi K2.5 vs Qwen3.5-35B-A3B

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 7,000+ subscribers.

Last updated November 19, 2026