Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Moonshot AI / Kimi / Kimi research agents

Kimi-Researcher

Research agent trained end to end with reinforcement learning on an internal Kimi k-series model for multi-step search and reasoning.

Language models (LLMs)
restrictedClosed weightsText generationReasoningAgents
Official documentation (opens in a new tab)Reviewed 14 Oct 2026
5 of 17 specification fields collected
Context windowNot collectedAwaiting source review
Maximum outputNot collectedAwaiting source review
Input billingNot collectedMoonshot research
Output billingNot collectedNo inferred price

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
Not collected
Maximum output
Not collected
Input modalities
TextSource for Input modalities (opens in a new tab)
Output modalities
TextSource for Output modalities (opens in a new tab)
Reasoning
SupportedSource for Reasoning (opens in a new tab)
Tool calling
SupportedSource for Tool calling (opens in a new tab)

Tool use inside the research system, not a documented public tools API.

Structured output
Not collected
Streaming
Not collected
Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
Not collected
Parameters
Not collected
Architecture
Not collected
Languages
Not collected
Context caching
Not collected
Published training method
End-to-end agentic reinforcement learning, primarily REINFORCE, with context management and turn-level partial rolloutsSource for Published training method (opens in a new tab)

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Moonshot research

Provider source (opens in a new tab)
API model ID
Not collected

Research identifier, not a verified public API selector.

Access
Research release. No current public model endpoint, model-weight download or price is established for this exact variant.
Regions
Historical native platform or research access only; no current regional availability is asserted.

Endpoint pricing has not been collected. Check the provider for applicable rates.

Data & enterprise terms

Retention, training use, regional processing and enterprise terms have not been summarized for this endpoint.

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Moonshot research

Not verified
How it works
Not collected
Minimum reusable input
Not collected
Lifetime & refresh
Not collected
How to enable it
Not collected
Usage fields
Not collected
Billing rules
Not collected
Limits & data boundaries
Not collected

No verified cache rate for this endpoint. An empty rate is not a zero charge.

What the measurements say

Humanity's Last Exam Pass@1

Moonshot AI · provider-reported
26.9%

Provider evaluation on June 17, 2025 using o3-mini as evaluator and research-system tools. Search-engine changes can affect results; not a current API benchmark.

Evaluation source (opens in a new tab)

Humanity's Last Exam Pass@4

Moonshot AI · provider-reported
40.17%

Provider-reported four-attempt accuracy; do not compare as single-attempt Pass@1. Evaluated June 17, 2025 with o3-mini and research-system tools.

Evaluation source (opens in a new tab)

xbench-DeepSearch Pass@1

Moonshot AI · provider-reported
69%

Provider evaluation June 18, 2025, average of four runs, o3-mini evaluator. Tool/search fluctuations affect reproducibility.

Evaluation source (opens in a new tab)

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

The developer identifies a specifically trained internal k-series model and describes a future plan to release its weights. No exact Kimi-Researcher checkpoint was found in the reviewed public inventory. This is a historical research-system record, not a claim that all later Kimi agents are closed-weight.

Classification evidence (opens in a new tab)

Lifecycle dates

Research announcement20 Jun 2025Source (opens in a new tab)

Sources reviewed