ai.canshu.com
中文
VRAM Calculator Advisor & TCO API Pricing Compare Models Leaderboard GPUs

Search models, GPUs, and tools

↑↓ Navigate · Enter Open · ★ Watchlist prioritized ai.canshu.com
Moonshot Kimi
Open-Weight Moonshot Kimi · MoE 1.03T / A32B · Text+Image+Video · BF16 / FP8

Kimi K2.5 (1T MoE) Architecture Parameters, Benchmarks & VRAM

1T total / 32B active multimodal MoE supporting Agent Swarm orchestration.

🧬 Kimi K2.5 (1T MoE) Architecture & Tensor Specifications

Verified directly against HuggingFace (moonshotai/Kimi-K2.5) config.json & Model Card

HF Class: KimiK25ForConditionalGeneration

1. Backbone & Dimensions

kimi_k25
Total / Active Params
1.026T / 32B
Hidden Size (d_model)
7,168
FFN Intermediate Size
18,432
MoE Expert Dim
2,048
MoE Experts
384 Routed · Top-8 + 1 Shared
Activation / Norm
SiLU (SwiGLU) · ε=0.00001
Native Tensor Dtype
BF16 / FP8

2. Attention & Layer Layout

MLA
Num Hidden Layers
61 Layers
Hidden Layer Layout
61-Layer Sparse MoE (384 Experts, Top-8 Active)
Attention Heads (Q / KV)
64 Q / 64 KV
Head Dimension
112

3. Context, RoPE & Tokenizer

Tokenizer
Native Context Length
262,144 tokens
RoPE Configuration
RoPE (原生 262,144 Context)
Vocab Size
163,840 (Untied)

📋 Kimi K2.5 (1T MoE) Official API & System Card Specifications

Verified against official Moonshot Kimi developer documentation & pricing

API Model ID: kimi-k2.5

1. Identity & Capabilities

API Spec
API Model ID
kimi-k2.5
Release Date
2026-01-27
Weight License
Open-Weight (API Spec)
Native Reasoning
✅ Supported
Function Calling & JSON Schema
部分支持
Batch API 异步批处理
✅ Supported (50% off)

2. Context, Output & Modalities

Limits
Max Context Window
256,000 tokens (256K)
Max Output Tokens
33,000 tokens
Input Modalities
Text + Image + Video
Output Modalities
Text
Direct China Access
✅ Direct CN + RMB
Deployment Mode
Open-Weight + Cloud API

3. Official Pricing & Cache

CNY / 1M
Input / 1M Tokens
¥ 4
Cached Input / 1M
¥ 0.8 -80%
Output / 1M Tokens
¥ 21
Tiered Pricing Notes
Alibaba Bailian: Hosted on Alibaba Cloud Bailian (CN Direct)

🎛️ Official Sampling & Reasoning Parameters

Recommended generation settings & API parameter constraints

Official Default Sampling
temperature=0.7 top_p=0.95 max_tokens=32,768

📊 Kimi K2.5 (1T MoE) Full Benchmark Scorecard

39 public benchmarks · ranks among 154 models · synced 2026-11-20

⚔️ Compare head-to-head

💻 Coding 8 items

SWE-Bench Verified #25/59 76.8%

A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.

SWE-bench Multilingual #19/33 73%

A multilingual benchmark for issue resolving in software engineering that covers Java, TypeScript, JavaScript, Go, Rust, C, and C++.

PaperBench #2/3 63.5%

PaperBench is a benchmark for evaluating AI agents on their ability to replicate research papers. It tests models on complex, multi-step workflows involving code implementation, experimentation, and reproducing scientific results from academic publications.

Terminal-Bench 2.0 #27/38 50.8%

Terminal-Bench 2.0 is an updated benchmark for testing AI agents' tool use ability to operate a computer via terminal.

SWE-Bench Pro #48/52 50.7%

SWE-Bench Pro is an advanced version of SWE-Bench that evaluates language models on complex, real-world software engineering tasks requiring extended reasoning and multi-step problem solving.

SciCode #8/24 48.7%

SciCode is a research coding benchmark curated by scientists that challenges language models to code solutions for scientific problems.

CyberGym #14/15 41.3%

CyberGym is a benchmark for evaluating AI agents on cybersecurity tasks, testing their ability to identify vulnerabilities, perform security analysis, and complete security-related challenges in a controlled environment.

FrontierSWE #14/16 26%

FrontierSWE measures whether an agent can complete open-ended technical projects at the scale of hours to tens of hours, spanning systems optimization, large-scale code construction, and applied ML research. Performance is reported as a dominance score, where higher is better.

📐 Math 8 items

AIME 2025 #6/29 96.1%

All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999.

HMMT 2025 #2/11 95.4%

Harvard-MIT Mathematics Tournament 2025 - A prestigious student-organized mathematics competition for high school students featuring two tournaments (November 2025 at MIT and February 2026 at Harvard) with individual tests, team rounds, and guts rounds.

MathVista-Mini #1/7 90.1%

MathVista-Mini is a smaller version of the MathVista benchmark that evaluates mathematical reasoning in visual contexts.

MMLU-Pro #7/39 87.1%

A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks.

MathVision #13/20 84.2%

MATH-Vision is a dataset designed to measure multimodal mathematical reasoning capabilities.

IMO-AnswerBench #14/17 81.8%

IMO-AnswerBench is a benchmark for evaluating mathematical reasoning capabilities on International Mathematical Olympiad (IMO) problems, focusing on answer generation and verification.

LiveBench #24/27 69.1%

LiveBench is a challenging, contamination-limited LLM benchmark that addresses test set contamination by releasing new questions monthly based on recently-released datasets, arXiv papers, news articles, and IMDb movie synopses.

Humanity's Last Exam #21/69 50.2%

Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions.

👁️ Multimodal 14 items

InfoVQAtest #1/1 92.6%

InfoVQA test set with infographic images requiring joint reasoning over document layout, textual content, graphical elements, and data visualizations with elementary reasoning and arithmetic skills.

OCRBench #1/5 92.3%

OCRBench: Comprehensive evaluation benchmark for assessing Optical Character Recognition (OCR) capabilities in Large Multimodal Models across text recognition, scene text VQA, and document understanding tasks.

OmniDocBench 1.5 #10/16 88.8%

OmniDocBench 1.5 is a comprehensive benchmark for evaluating multimodal large language models on document understanding tasks, including OCR, document parsing, information extraction, and visual question answering across diverse document types.

Video-MME #4/8 87.4%

Video-MME is the first-ever comprehensive evaluation benchmark of Multi-modal Large Language Models (MLLMs) in video analysis.

VideoMMMU #2/14 86.6%

Video-MMMU evaluates Large Multimodal Models' ability to acquire knowledge from expert-level professional videos across six disciplines through three cognitive stages: perception, comprehension, and adaptation.

MMVU #2/7 80.4%

MMVU (Multimodal Multi-disciplinary Video Understanding) is a benchmark for evaluating multimodal models on video understanding tasks across multiple disciplines, testing comprehension and reasoning capabilities on video content.

LongVideoBench #2/4 79.8%

LongVideoBench is a question-answering benchmark featuring video-language interleaved inputs up to an hour long.

MMMU-Pro #16/44 78.5%

A more robust multi-discipline multimodal understanding benchmark that enhances MMMU through a three-step process: filtering text-only answerable questions, augmenting candidate options, and introducing vision-only input settings.

CharXiv-R #30/41 77.5%

CharXiv-R is the reasoning component of the CharXiv benchmark, focusing on complex reasoning questions that require synthesizing information across visual chart elements.

LVBench #9/15 75.9%

LVBench is an extreme long video understanding benchmark designed to evaluate multimodal models on videos up to two hours in duration. It contains 6 major categories and 21 subcategories, with videos averaging five times longer than existing datasets.

MotionBench #3/4 70.4%

MotionBench is a benchmark for evaluating multimodal models on motion understanding in videos, testing the ability to comprehend temporal dynamics, movement patterns, and action sequences.

WorldVQA #6/6 46.3%

WorldVQA is a benchmark designed to evaluate atomic vision-centric world knowledge. It assesses models' ability to understand and reason about visual elements representing real-world knowledge.

SimpleVQA #6/13 0.7

SimpleVQA is a visual question answering benchmark focused on simple queries.

ZEROBench #9/13 0.1

ZEROBench is a challenging vision benchmark designed to test models on zero-shot visual understanding tasks.

Source: ZeroEval / llm-stats.com (✓ = third-party verified, others vendor-reported).

⚡ Kimi K2.5 (1T MoE) Interactive VRAM & Token Speed Calculator

Concurrency (Batch):
MoE 1026B / A32B·61L (d=7168)·64Q/64KV·Ctx 256K·Text+Image+Video
Runnable Models →
GPUs
96 GB VRAM·4000 GB/s·Safe Usable 94.1 GB
Out of VRAM (OOM)636.9 / 94.1 GBOOM
3. Weight Quantization
4. Per-Request Context16K (16,384 tok)
Verified: 2026-11-20·Roofline bandwidth/compute estimate (Batch=1)
Out of VRAM (OOM)
636.9/ 94.1 GB usable

Physical 96 GB · Load 677%

Theoretical Decode Speed
—tokens/s

Exceeds usable VRAM by 542.82 GB

Max Context at Batch=1
0K/ 256K

MLA compresses multi-batch KV footprint

Weights: 622.01 GBKV Cache (1×16K): 1.15 GBOverhead: 13.74 GB
⛔ Cannot Load
Reference Launch CLI:
OLLAMA_CONTEXT_LENGTH=16384 ollama run kimi-k2-5-1000b

Insufficient Local VRAM? Ways to Run Kimi K2.5 (1T MoE)

Rent an hourly high-VRAM cloud GPU pod, or check the 12-month Buy vs Cloud vs API break-even curve.

📊 Kimi K2.5 (1T MoE) Quantization VRAM Breakdown (16K Context)

Quantization Bits/Weight Weights KV Cache (16K) Total VRAM Quality Note
FP16 / BF16 16 bpw 2052 GB 1.15 GB 2095.49 GB Lossless Full Precision
FP8 / Q8_0 8.5 bpw 1090.13 GB 1.15 GB 1114.38 GB 99.9% Near-Lossless
GGUF Q6_K 6.56 bpw 841.32 GB 1.15 GB 860.6 GB Ultra High Quality
GGUF Q5_K_M 5.69 bpw 729.74 GB 1.15 GB 746.79 GB High Balanced Quality
Q4_K_M / AWQ 4.85 bpw 622.01 GB 1.15 GB 636.9 GB Sweet Spot (Recommended)
GGUF IQ4_XS 4.25 bpw 545.06 GB 1.15 GB 558.42 GB Compact 4-bit
GGUF Q3_K_M 3.91 bpw 501.46 GB 1.15 GB 513.94 GB Moderate Perplexity Loss
GGUF Q2_K 3.15 bpw 403.99 GB 1.15 GB 414.52 GB Extreme Compression

🖥️ GPU & Apple Silicon Compatibility for Kimi K2.5 (1T MoE) (Q4_K_M)

Frequently Asked Questions

What are the exact architecture parameters of Kimi K2.5 (1T MoE)?

Kimi K2.5 (1T MoE) uses KimiK25ForConditionalGeneration with 61 layers (61-Layer Sparse MoE (384 Experts, Top-8 Active)), hidden_size=7168, 64 Q heads / 64 KV heads (head_dim=112), and vocab_size=163,840.

How much VRAM does Kimi K2.5 (1T MoE) require?

At 4-bit quantization (Q4_K_M, 16K context), Kimi K2.5 (1T MoE) requires 636.9 GB VRAM. Full FP16 requires 2095.49 GB.