Skip to main content
BenchLM
Data

Kimi K3.1 Pro vs d1.2

Updated November 17, 2026. We do not rank this pair: at least one has no public score. Public scores include evidence status and uncertainty.

Share or export
Share on XLinkedInAPI/MCP

Decision reading

The public evidence has no benchmark result shared by both models, so it does not support a quality verdict. Use the documented cost, context, and runtime rows instead. 0 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

71.04/100

Estimated · Public rank #17

Conditional range 56.7–85.4

Model B
Liquid AI logo

Liquid AI

—

Evidence status unavailable

90% interval unavailable

API price per 1M tokens

Kimi K3.1 Pro: $2 input / $10 output

d1.2: $0.020 input / $0 output

Context window

Kimi K3.1 Pro: 1.05M

d1.2: Not sourced

Released

Kimi K3.1 Pro: October 23, 2026

d1.2: November 16, 2026

Shared results
0
Kimi K3.1 Pro only
49
d1.2 only
0
Like-for-like categories
0 / 8
Estimated: Kimi K3.1 Pro. 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.

  • Chat turn cost

    1K fresh input + 500 output tokens

    d1.2

    d1.2 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.

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

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

    d1.2

    d1.2 has the lower estimated token cost for this stated workload. d1.2 has no published cached-input rate, so cached tokens use its listed input rate.

    Confidence: rate-fallback
Show secondary and unsupported calls
  • Repository review cost

    50K fresh input + 3K output tokens

    d1.2

    d1.2 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.

    Confidence: listed-rates
  • Coding work

    Code generation, repair, and software-engineering tasks

    Not enough matched evidence

    D1.2 is not ranked on the public lane for coding, so no winner is named for coding.

    Confidence: limited
  • Agentic work

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

    Not enough matched evidence

    D1.2 is not ranked on the public lane for agentic, so no winner is named for agentic.

    Confidence: limited
  • Long documents

    Prompts that approach the documented context limit

    Not enough matched evidence

    A complete context comparison is not sourced.

    Confidence: limited

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.

61.1Kimi K3.1 Pro—d1.2

Not comparable · BenchAlign v5.8

The coding row is not comparable on the public lane: at least one model is not measured or not ranked there.

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.

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.

A shared-evidence shape is not available.

BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.

Category results, on a stated basis

Each row shows the public-lane category score for both models: the BenchAlign v5.8 lane for agentic, coding, reasoning, multimodal, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.

Agentic

Not comparable
Kimi K3.1 Pro
70.6
Supported · #13/138
d1.2
Not ranked
Basis
BenchAlign v5.8 lane · 20 vs 0 public rows
Reading
Not comparable

Coding

Not comparable
Kimi K3.1 Pro
61.1
Supported · #17/163
d1.2
Not ranked
Basis
BenchAlign v5.8 lane · 9 vs 0 public rows
Reading
Not comparable

Reasoning

Not comparable
Kimi K3.1 Pro
69.2
Supported · #18/160
d1.2
Not ranked
Basis
BenchAlign v5.8 lane · 0 vs 0 public rows
Reading
Not comparable

Multimodal

Not comparable
Kimi K3.1 Pro
71.8
Estimated · #12/67
d1.2
Not ranked
Basis
BenchAlign v5.8 lane · 15 vs 0 public rows
Reading
Not comparable

Knowledge

Not comparable
Kimi K3.1 Pro
74.1
Supported · #14/181
d1.2
Not ranked
Basis
BenchAlign v5.8 lane · 5 vs 0 public rows
Reading
Not comparable

Multilingual

Not comparable
Kimi K3.1 Pro
Not ranked
d1.2
Not ranked
Basis
Provisional lane · 0 vs 0 weighted rows
Reading
Not comparable

Instruction following

Not comparable
Kimi K3.1 Pro
Not ranked
d1.2
Not ranked
Basis
Provisional lane · 0 vs 0 weighted rows
Reading
Not comparable

Math

Not comparable
Kimi K3.1 Pro
Not ranked
d1.2
Not ranked
Basis
Provisional lane · 0 vs 0 weighted rows
Reading
Not comparable

Ranks count the models scored in each category’s lane, so the agentic, coding, reasoning, multimodal, and knowledge denominators (BenchAlign v5.8) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. 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 K3.1 Pro
$0.007
Fits in one request
d1.2
$0.00002
Fit state unavailable

d1.2 has the lower modeled cost

Costs use the listed standard API rates.

Repository review

50K fresh input + 3K output tokens

Kimi K3.1 Pro
$0.13
Fits in one request
d1.2
$0.001
Fit state unavailable

d1.2 has the lower modeled cost

Costs use the listed standard API rates.

Cache-heavy agent loop

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

Kimi K3.1 Pro
$0.18
Fits in one request
d1.2
$0.0044
Fit state unavailable
Cached input priced at the published list-input rate

d1.2 has the lower modeled cost

d1.2 has no published cached-input rate, so cached tokens use its listed input rate.

Your workload

Enter a month of tokens to price it at both models’ listed API rates.

Kimi K3.1 Pro
$40.00
per month
d1.2
$0.200
per month

Standard listed rates, no batch or volume discounts.

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 17, 2026.

Kimi K3.1 Pro

20406080100$0.10$0.50$1$5$10$50$100Output price, $ per 1M tokens (log)Overall score
  • 2 models at the same price score higher than Kimi K3.1 Pro. The nearest is Gemini 4 Carbon (score 89.6, also $10.00 per 1M output tokens).
  • d1.2 has no overall score yet, 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 K3.1 Pro 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 K3.1 Pro

1.05M

d1.2

Cached-input rate

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

Kimi K3.1 Pro

$0.2 per 1M cached input tokens

d1.2

Not published

Liquid AI pricing

Provider availability

Kimi K3.1 Pro

Not sourced

d1.2

Generally Available · Liquid API, Liquid Console

Liquid AI d1.2 launch

Reasoning profile

Kimi K3.1 Pro

Reasoning

d1.2

Non-Reasoning

Weight access

Kimi K3.1 Pro

Open Weight

d1.2

Proprietary

License

Kimi K3.1 Pro

Open Weight

d1.2

Proprietary

Release date

Kimi K3.1 Pro

2026-10-23

d1.2

2026-11-16

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
The public evidence has no benchmark result shared by both models, so it does not support a quality verdict.
Workload cost
Repository review: $0.13 vs $0.001. Cache-heavy agent loop: $0.18 vs $0.0044.
Context tradeoff
A complete documented context comparison is not available.
Run the same representative tasks against both endpoints before changing production traffic.

Questions

Which is better, Kimi K3.1 Pro or d1.2?

The public evidence has no benchmark result shared by both models, so it does not support a quality verdict. The page therefore keeps the decision tied to the specific documented workload.

Which is better for coding, Kimi K3.1 Pro or d1.2?

D1.2 is not ranked on the public lane for coding, so no winner is named for coding.

Which is better for agentic tasks, Kimi K3.1 Pro or d1.2?

D1.2 is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.

Which costs less, Kimi K3.1 Pro or d1.2?

For the stated presets, chat costs $0.007 on Kimi K3.1 Pro and $0.00002 on d1.2; repository review costs $0.13 and $0.001; the cache-heavy agent loop costs $0.18 and $0.0044. d1.2 has no published cached-input rate, so cached tokens use its listed input rate.

Which has the larger context window, Kimi K3.1 Pro or d1.2?

A complete documented context-window comparison is not available.

Self-host vs API cost

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

Kimi K3.1 Pro
API / mo$9,000
Self-host / mo$57,606
Break-evenNot reached at 70% load
d1.2
API / mo$15
Self-host / moNot listed
Break-even—
Proprietary model — self-hosting not applicable.
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 evidence49 rows

Agentic

  • Terminal-Bench 4.1

    Kimi K3.1 Pro59.70%
    Source
    d1.2—

    Not directly comparable

  • Terminal-Bench 4.0

    Kimi K3.1 Pro65.70%
    Source
    d1.2—

    Not directly comparable

  • Terminal-Bench-Science 0.1

    Kimi K3.1 Pro38.1%
    Source
    d1.2—

    Not directly comparable

  • HLE w/ tools

    Kimi K3.1 Pro70.0%
    Source
    d1.2—

    Not directly comparable

  • BrowseComp

    Kimi K3.1 Pro92.1%
    Source
    d1.2—

    Not directly comparable

  • BioMysteryBench

    Kimi K3.1 Pro80.6%
    Source
    d1.2—

    Not directly comparable

  • Agents' Last Exam

    Kimi K3.1 Pro43.2%
    Source
    d1.2—

    Not directly comparable

  • Agents' Last Exam (CLI)

    Kimi K3.1 Pro44.2%
    Source
    d1.2—

    Not directly comparable

  • APEX-Agents

    Kimi K3.1 Pro69.4%
    Source
    d1.2—

    Not directly comparable

  • Frontier Finance

    Kimi K3.1 Pro58.6%
    Source
    d1.2—

    Not directly comparable

  • HANDBOOK.md

    Kimi K3.1 Pro44.6%
    Source
    d1.2—

    Not directly comparable

  • OSWorld 2.0

    Kimi K3.1 Pro79.6%
    Source
    d1.2—

    Not directly comparable

  • Business Operation

    Kimi K3.1 Pro37.8%
    Source
    d1.2—

    Not directly comparable

  • Financial Modeling

    Kimi K3.1 Pro57.6%
    Source
    d1.2—

    Not directly comparable

  • Financial Research

    Kimi K3.1 Pro70.4%
    Source
    d1.2—

    Not directly comparable

  • Kimi Hallu Bench Hard (tools)

    Kimi K3.1 Pro8.3%
    Source
    d1.2—

    Not directly comparable

  • ExploitGym

    Kimi K3.1 Pro20.7%
    Source
    d1.2—

    Not directly comparable

  • MazeBench

    Kimi K3.1 Pro2%
    Source
    d1.2—

    Not directly comparable

  • OSWorld 2.1

    Kimi K3.1 Pro79.6%
    Source
    d1.2—

    Not directly comparable

  • CWE-bench

    Kimi K3.1 Pro45.3%
    Source
    d1.2—

    Not directly comparable

Coding

  • DeepSWE

    Kimi K3.1 Pro75.5%
    Source
    d1.2—

    Not directly comparable

  • ProgramBench (binary)

    Kimi K3.1 Pro31.5%
    Source
    d1.2—

    Not directly comparable

  • Kimi Code Bench v2

    Kimi K3.1 Pro77.1%
    Source
    d1.2—

    Not directly comparable

  • Kimi Code Bench 3.0 Long-horizon

    Kimi K3.1 Pro61.2%
    Source
    d1.2—

    Not directly comparable

  • sweMarathon

    Kimi K3.1 Pro51.7%
    Source
    d1.2—

    Not directly comparable

  • PostTrainBench

    Kimi K3.1 Pro36.5%
    Source
    d1.2—

    Not directly comparable

  • Bug Hunt Bench

    Kimi K3.1 Pro36.0 fixes
    Source
    d1.2—

    Not directly comparable

  • FrontierCode 2.0

    Kimi K3.1 Pro19.9%
    Source
    d1.2—

    Not directly comparable

  • OpenHarmony Bench v1.2

    Kimi K3.1 Pro52.38%
    Source
    d1.2—

    Not directly comparable

Multimodal

  • OfficeQA Pro

    Kimi K3.1 Pro67.4%
    Source
    d1.2—

    Not directly comparable

  • OfficeQA Pro v2

    Kimi K3.1 Pro64.4%
    Source
    d1.2—

    Not directly comparable

  • BenchCAD Vision2Code (no tools)

    Kimi K3.1 Pro0.728
    Source
    d1.2—

    Not directly comparable

  • BenchCAD Vision2Code (tools)

    Kimi K3.1 Pro0.927
    Source
    d1.2—

    Not directly comparable

  • MMMU-Pro

    Kimi K3.1 Pro84.1%
    Source
    d1.2—

    Not directly comparable

  • MMMU-Pro w/ Python

    Kimi K3.1 Pro90%
    Source
    d1.2—

    Not directly comparable

  • CharXiv

    Kimi K3.1 Pro93.4%
    Source
    d1.2—

    Not directly comparable

  • CharXiv w/o tools

    Kimi K3.1 Pro89%
    Source
    d1.2—

    Not directly comparable

  • Chartography (tools)

    Kimi K3.1 Pro90.2%
    Source
    d1.2—

    Not directly comparable

  • BabyVision w/ Python

    Kimi K3.1 Pro95.4%
    Source
    d1.2—

    Not directly comparable

  • ZeroBench

    Kimi K3.1 Pro39.0%
    Source
    d1.2—

    Not directly comparable

  • ZeroBench w/ Python

    Kimi K3.1 Pro63.0%
    Source
    d1.2—

    Not directly comparable

  • GDP.pdf (Gemini judge)

    Kimi K3.1 Pro34.6%
    Source
    d1.2—

    Not directly comparable

  • GDP.pdf w/ Python (Gemini judge)

    Kimi K3.1 Pro85.7%
    Source
    d1.2—

    Not directly comparable

  • PerceptionBench

    Kimi K3.1 Pro68.3%
    Source
    d1.2—

    Not directly comparable

Knowledge

  • GPQA

    Kimi K3.1 Pro94.2%
    Source
    d1.2—

    Not directly comparable

  • GPQA-D

    Kimi K3.1 Pro94.2%
    Source
    d1.2—

    Not directly comparable

  • HLE

    Kimi K3.1 Pro70%
    Source
    d1.2—

    Not directly comparable

  • HLE w/o tools

    Kimi K3.1 Pro51.1%
    Source
    d1.2—

    Not directly comparable

  • Kimi Hallu Bench

    Kimi K3.1 Pro5.4%
    Source
    d1.2—

    Not directly comparable

49 public results · 0 shared

Watch Kimi K3.1 Pro vs d1.2

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 17, 2026