Skip to main content
Radar

Provider changes are easy to miss. Radar watches releases, pricing, deprecations, and incidents at the source.Provider changes are easy to miss.

See Radar
BenchLM
Advertise

Model comparison

Exaone 4.0 32B vs Kimi K2.5

Updated August 7, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

Exaone 4.0 32B

LG AI Research

39.5/100

Estimated · Public rank #186

90% interval 28.0–51.1

Kimi K2.5

Moonshot AI

58.8/100

Supported · Public rank #64

90% interval 51.4–66.2

Kimi K2.5 has the higher public score, 58.82 versus 39.55, and the 90% score intervals do not overlap.

2 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.

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.

  • Long documents

    Prompts that approach the documented context limit

    Kimi K2.5

    Kimi K2.5 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

    No shared weighted benchmark basis supports a winner.

    Confidence: limited

  • Agentic work

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

    Not enough matched evidence

    No shared weighted benchmark basis supports a winner.

    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

    The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. Exaone 4.0 32B does not fit this workload in one request. Kimi K2.5 has no published cached-input rate, so cached tokens use its listed input rate. Exaone 4.0 32B has no comparable published API token rate.

    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

What is actually comparable

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

Shared results
2
Exaone 4.0 32B only
0
Kimi K2.5 only
43
Like-for-like categories
0 / 8

1 category uses different evidence sets. Those rows remain visible for coverage context but do not name a winner.

Category results, on a stated basis

Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.

Knowledge

Directional only
Exaone 4.0 32B
81.8
Kimi K2.5
56.9
Weighted basis
1 vs 4 rows
Reading
Directional only

Agentic

Not comparable
Exaone 4.0 32B
Not measured
Kimi K2.5
55.0
Weighted basis
0 vs 2 rows
Reading
Not comparable

Coding

Not comparable
Exaone 4.0 32B
Not measured
Kimi K2.5
59.4
Weighted basis
0 vs 4 rows
Reading
Not comparable

Reasoning

Not comparable
Exaone 4.0 32B
Not measured
Kimi K2.5
61.0
Weighted basis
0 vs 1 rows
Reading
Not comparable

Math

Not comparable
Exaone 4.0 32B
Not measured
Kimi K2.5
60.6
Weighted basis
0 vs 4 rows
Reading
Not comparable

Multilingual

Not comparable
Exaone 4.0 32B
Not measured
Kimi K2.5
82.3
Weighted basis
0 vs 1 rows
Reading
Not comparable

Multimodal

Not comparable
Exaone 4.0 32B
Not measured
Kimi K2.5
78.5
Weighted basis
0 vs 1 rows
Reading
Not comparable

Instruction following

Not comparable
Exaone 4.0 32B
Not measured
Kimi K2.5
93.9
Weighted basis
0 vs 1 rows
Reading
Not comparable

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.

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

Exaone 4.0 32B
API rate not published
Fits in one request
Kimi K2.5
$0.0021
Fits in one request

Exaone 4.0 32B has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

Exaone 4.0 32B
API rate not published
Fits in one request
Kimi K2.5
$0.039
Fits in one request

Exaone 4.0 32B has no comparable published API token rate.

Cache-heavy agent loop

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

Exaone 4.0 32B
API rate not published
Does not fit in one request
Cached-input rate unavailable
Kimi K2.5
$0.162
Fits in one request
Cached input priced at the published list-input rate

Exaone 4.0 32B does not fit this workload in one request. Kimi K2.5 has no published cached-input rate, so cached tokens use its listed input rate. Exaone 4.0 32B has no comparable published API token rate.

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.

Exaone 4.0 32B

128K

Kimi K2.5

256K

API model ID

Exaone 4.0 32B

Not sourced

Kimi K2.5

Not sourced

Cached-input rate

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

Exaone 4.0 32B

No comparable hosted API rate

Kimi K2.5

Not published

Documented inputs

Exaone 4.0 32B

Not sourced

Kimi K2.5

Not sourced

Documented outputs

Exaone 4.0 32B

Not sourced

Kimi K2.5

Not sourced

Provider availability

Exaone 4.0 32B

Not sourced

Kimi K2.5

Not sourced

Reasoning profile

Exaone 4.0 32B

Reasoning

Kimi K2.5

Non-Reasoning

Weight access

Exaone 4.0 32B

Open Weight

Kimi K2.5

Open Weight

License

Exaone 4.0 32B

Open Weight

Kimi K2.5

Open Weight

Release date

Exaone 4.0 32B

Not sourced

Kimi K2.5

2026-02-01

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 score, 58.82 versus 39.55, and the 90% score intervals do not overlap.
Workload cost
A complete comparable API-rate estimate is not available for both models.
Context tradeoff
Kimi K2.5 has the larger documented window (256K).

Run the same representative tasks against both endpoints before changing production traffic.

Self-host vs API cost

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

Exaone 4.0 32B
API / mo$0
Self-host / moNot listed
Break-even—
Proprietary model — self-hosting not applicable.
Kimi K2.5
API / mo$2,700
Self-host / mo$5,221
Break-even132M/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 evidence45 rows

Agentic

  • Terminal-Bench 2.0

    Exaone 4.0 32B—
    Kimi K2.550.8%
    Source

    Not directly comparable

  • BrowseComp

    Exaone 4.0 32B—
    Kimi K2.560.6%
    Source

    Not directly comparable

  • Claw-Eval

    Exaone 4.0 32B—
    Kimi K2.552.3%
    Source

    Not directly comparable

  • QwenClawBench

    Exaone 4.0 32B—
    Kimi K2.554.3%
    Source

    Not directly comparable

  • τ³-bench results

    Exaone 4.0 32B—
    Kimi K2.565.7%
    Source

    Not directly comparable

  • DeepSearchQA

    Exaone 4.0 32B—
    Kimi K2.577.1%
    Source

    Not directly comparable

  • DeepPlanning

    Exaone 4.0 32B—
    Kimi K2.514.4%
    Source

    Not directly comparable

  • Toolathlon

    Exaone 4.0 32B—
    Kimi K2.527.8%
    Source

    Not directly comparable

  • MCP Atlas

    Exaone 4.0 32B—
    Kimi K2.529.5%
    Source

    Not directly comparable

  • MCP-Tasks

    Exaone 4.0 32B—
    Kimi K2.559.1%
    Source

    Not directly comparable

  • WideResearch

    Exaone 4.0 32B—
    Kimi K2.572.7%
    Source

    Not directly comparable

  • Gert Labs

    Exaone 4.0 32B—
    Kimi K2.545.88%
    Source

    Not directly comparable

  • ResearchClawBench

    Exaone 4.0 32B—
    Kimi K2.514.0%
    Source

    Not directly comparable

  • JobBench

    Exaone 4.0 32B—
    Kimi K2.58.7%
    Source

    Not directly comparable

Coding

  • SWE-bench Verified

    Exaone 4.0 32B—
    Kimi K2.576.8%
    Source

    Not directly comparable

  • SWE-bench Verified*

    Exaone 4.0 32B—
    Kimi K2.570.8%
    Source

    Not directly comparable

  • LiveCodeBench v6

    Exaone 4.0 32B—
    Kimi K2.585.0%
    Source

    Not directly comparable

  • SWE-bench Pro

    Exaone 4.0 32B—
    Kimi K2.550.7%
    Source

    Not directly comparable

  • SWE Multilingual

    Exaone 4.0 32B—
    Kimi K2.573%
    Source

    Not directly comparable

  • SWE-Rebench

    Exaone 4.0 32B—
    Kimi K2.558.5%
    Source

    Not directly comparable

  • React Native Evals

    Exaone 4.0 32B—
    Kimi K2.577.2%
    Source

    Not directly comparable

  • SciCode

    Exaone 4.0 32B—
    Kimi K2.548.7%
    Source

    Not directly comparable

Reasoning

  • LongBench v2

    Exaone 4.0 32B—
    Kimi K2.561%
    Source

    Not directly comparable

Knowledge

  • MMLU-Pro

    Exaone 4.0 32B81.8%
    Source
    Kimi K2.587.1%
    Source

    Kimi K2.5 leads this result

  • GPQA

    Exaone 4.0 32B—
    Kimi K2.587.6%
    Source

    Not directly comparable

  • GPQA-D

    Exaone 4.0 32B—
    Kimi K2.587.6%
    Source

    Not directly comparable

  • SuperGPQA

    Exaone 4.0 32B—
    Kimi K2.569.2%
    Source

    Not directly comparable

  • MMLU-Pro (Arcee)

    Exaone 4.0 32B—
    Kimi K2.587.1%
    Source

    Not directly comparable

  • HLE

    Exaone 4.0 32B—
    Kimi K2.530.1%
    Source

    Not directly comparable

Math

  • AIME 2025

    Exaone 4.0 32B85.3%
    Source
    Kimi K2.596.1%
    Source

    Kimi K2.5 leads this result

  • AIME26

    Exaone 4.0 32B—
    Kimi K2.595.8%
    Source

    Not directly comparable

  • AIME25 (Arcee)

    Exaone 4.0 32B—
    Kimi K2.596.3%
    Source

    Not directly comparable

  • HMMT Feb 2025

    Exaone 4.0 32B—
    Kimi K2.595.4%
    Source

    Not directly comparable

  • HMMT Nov 2025

    Exaone 4.0 32B—
    Kimi K2.591.1%
    Source

    Not directly comparable

  • HMMT Feb 2026

    Exaone 4.0 32B—
    Kimi K2.587.1%
    Source

    Not directly comparable

  • MMAnswerBench

    Exaone 4.0 32B—
    Kimi K2.581.8%
    Source

    Not directly comparable

  • FrontierMath v2 (Tiers 1-3)

    Exaone 4.0 32B—
    Kimi K2.527.900%
    Source

    Not directly comparable

  • FrontierMath v2 (Tier 4)

    Exaone 4.0 32B—
    Kimi K2.54.200%
    Source

    Not directly comparable

Multilingual

  • MMLU-ProX

    Exaone 4.0 32B—
    Kimi K2.582.3%
    Source

    Not directly comparable

  • NOVA-63

    Exaone 4.0 32B—
    Kimi K2.556.0%
    Source

    Not directly comparable

Multimodal

  • MMMU-Pro

    Exaone 4.0 32B—
    Kimi K2.578.5%
    Source

    Not directly comparable

  • Video-MME

    Exaone 4.0 32B—
    Kimi K2.587.4%
    Source

    Not directly comparable

  • MMVU

    Exaone 4.0 32B—
    Kimi K2.580.4%
    Source

    Not directly comparable

  • VideoMMMU

    Exaone 4.0 32B—
    Kimi K2.586.6%
    Source

    Not directly comparable

Instruction following

  • IFEval

    Exaone 4.0 32B—
    Kimi K2.593.9%
    Source

    Not directly comparable

Frequently asked questions

Which is better, Exaone 4.0 32B or Kimi K2.5?

Kimi K2.5 has the higher public score, 58.82 versus 39.55, and the 90% score intervals do not overlap. That is the clearest overall quality signal in the current public data.

Which is better for coding, Exaone 4.0 32B or Kimi K2.5?

The published evidence does not provide a shared weighted coding basis for both models, so BenchLM does not name a coding winner.

Which is better for agentic tasks, Exaone 4.0 32B or Kimi K2.5?

The published evidence does not provide a shared weighted agentic tasks basis for both models, so BenchLM does not name a agentic tasks winner.

Which costs less, Exaone 4.0 32B or Kimi K2.5?

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, Exaone 4.0 32B or Kimi K2.5?

Kimi K2.5 has the larger documented context window: 256K, compared with 128K.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated August 7, 2026

Watch Exaone 4.0 32B vs Kimi K2.5

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

Read a sample issue

Join 2,000+ readers.