LLM Reference

Kimi K2.5

Released
2026-03-15
Last refreshed
2026-06-29
Status
Researched 46d ago
ProprietaryCommercial use: conditionalMultimodalCodingRAGAgentsLong contextVisionClassificationJSON / Tool use

Kimi K2.5 is worth evaluating for coding, rag, and agents when its provider route and context window match the workload.

Use it for

  • Teams evaluating coding, rag, and agents
  • Workloads that can use a 256k context window
  • Buyers comparing 4 tracked provider routes

Do not use it for

  • Workloads where another current model has stronger sourced task evidence
Specifications
Family
Kimi
Released
2026-03-15
Context
256k
Parameters
1T (MoE, 384 experts)
Architecture
Mixture of Experts
Specialization
code
Openness
Proprietary
License
ProprietaryCommercial use: conditional
Weights
Not released
Code
Unknown
Training
Fine-tuned
Created by

Lossless long-context AI innovation

Beijing, China
Founded 2023
Website
Pricing
Output / 1M
$2.00
Input / 1M
$0.440

Cheapest of 10 routes · OpenRouter

About

Kimi K2.5 is Moonshot AI's Kimi model focused on code generation and software engineering. It offers a 256K-token context window and scores 87.9 on GPQA.

Kimi K2.5 is a proprietary model in the Kimi family. The structured metadata tracks a 256k-token context window, multimodal input, function calling, and structured outputs. This page tracks provider routes through Cloudflare Workers AI, Fireworks AI, OpenRouter, and 7 more, with the cheapest tracked route listed at $0.44 input and $2 output per 1M tokens. Headline tracked benchmarks include Google-Proof Q&A 87.9, MMLU PRO 87.1, and BFCL 47.1.

Top use-case fit: coding, agents, and build tasks

Coding

Q/$ C

2 relevant benchmarks in the decision map.

RAG

Included by capability and metadata signals in the decision map.

Agents

Q/$ C

4 relevant benchmarks in the decision map.

Provider price ladder

Compare all 10

Compare API pricing across 4 providers for input and output tokens, batch, and cached reads when available.

ProviderInput / 1MOutput / 1MRoute
OpenRouter$0.440$2.00
Serverless
Together AI$0.500$2.80
Serverless
AWS Bedrock$0.600$3.00
Serverless
Fireworks AI$0.600$3.00
Serverless

Available via routers & gateways(8)

Capabilities

VisionMultimodalFunction CallingStructured Outputs

Benchmark peer barsfor Coding

Benchmark scores(17)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionEvaluationSource
Google-Proof Q&A87.9diamondObserved 2026-04-18Source
MMLU PRO87.1Thinking mode (accuracy)Observed 2026-06-07Source
BFCL47.1v4Observed 2026-04-19Source
τ-bench74.2τ-benchObserved 2026-04-24Source
MultiChallenge61.4MultiChallengeObserved 2026-04-26Source
MMMU Pro78.5LLM-Stats aggregatorObserved 2026-06-07Source
SWE-rebench58.5pass@1 (best of 5 runs)Observed 2026-05-28Source
AIME 202596.1Thinking mode (accuracy)Observed 2026-06-07Source
Berkeley Function Calling Leaderboard v364.5BFCL v3 (accuracy%)Observed 2026-06-07Source
BrowseComp60.6BrowseComp (accuracy%)Observed 2026-06-07Source
Humanity's Last Exam50.2HLE-Full with tools (agentic) (accuracy)Observed 2026-06-07Source
LiveCodeBench85.0LiveCodeBench v6 (pass@1)Observed 2026-06-07Source
MATH-50098.0Thinking mode (accuracy)Observed 2026-06-07Source
MCP-Atlas29.5MCP-Atlas (accuracy%)Observed 2026-06-07Source
SWE-bench Verified76.8From official GitHub model card (resolved)Observed 2026-06-07Source
Terminal-Bench 2.050.8Terminal-Bench 2.0 (accuracy%)Observed 2026-06-07Source
CursorBench31.9CursorBench 3.1Observed 2026-06-30
Configuration: Kimi 2.5 (single reported configuration)
Harness: CursorBench 3.1
Evaluator: Cursor
Confidence: confirmed
Notes: Cursor published one CursorBench 3.1 configuration for this model; no cross-effort selection was needed.
Source

Migration checks

No linked migration route is available for this model yet.

Compare Kimi K2.5 with other models

Show all 65 popular comparisonssorted by 7-day search impressions
Kimi K2.5 vs GPT-5.5162Kimi K2.5 vs DeepSeek V3.1132Kimi K2.5 vs DeepSeek R1130Kimi K2.5 vs Xiaomi MiMo-V2.5128Kimi K2.5 vs Qwen3.5-397B-A17B123Kimi K2.5 vs Claude Opus 4.5105Kimi K2.5 vs Qwen3-235B-A22B105Kimi K2.5 vs GLM-5V-Turbo103Kimi K2.5 vs Gemini 3 Pro90Kimi K2.5 vs Xiaomi MiMo-V2.5-TTS-Series85Kimi K2.5 vs Qwen3.6-35B-A3B79Kimi K2.5 vs Qwen2.5-72B-Instruct73Kimi K2.5 vs Together AI Qwen2-72B-Instruct73Kimi K2.5 vs Claude Opus 4.672Kimi K2.5 vs Ling-2.6-Flash66Kimi K2.5 vs Qwen3.6-27B62Kimi K2.5 vs Grok-358Kimi K2.5 vs GLM-5 9B53Kimi K2.5 vs DeepSeek R1 Distill Llama 70B49Kimi K2.5 vs Mistral Large 3 675B Instruct47Kimi K2.5 vs GLM-5 Turbo46Kimi K2.5 vs DeepSeek R1 052846Kimi K2.5 vs Llama 3.1 70B Instruct45Kimi K2.5 vs Gemini 2.5 Flash Live API43Kimi K2.5 vs o338Kimi K2.5 vs Trinity-Large-Thinking38Kimi K2.5 vs Claude 3.7 Sonnet36Kimi K2.5 vs Together AI Qwen2-7B-Instruct33Kimi K2.5 vs Qwen3.5-35B-A3B32Kimi K2.5 vs GPT-5.227Kimi K2.5 vs Llama 3.1 405B Instruct27Kimi K2.5 vs Qwen2.5-7B-Instruct26Kimi K2.5 vs Llama 3 70B Instruct26Kimi K2.5 vs Gemini 2.5 Pro Computer Use Preview26Kimi K2.5 vs Llama 2 13B Chat24Kimi K2.5 vs Phi-3 Mini 4k23Kimi K2.5 vs Mistral Large 222Kimi K2.5 vs Tencent Hunyuan Turbo S21Kimi K2.5 vs Gemini 2.5 Flash20Kimi K2.5 vs Qwen3-9B20Kimi K2.5 vs Qwen3.5-27B19Kimi K2.5 vs Mistral Nemotron19Kimi K2.5 vs Gemma 7B Instruct17Kimi K2.5 vs o3 Mini17Kimi K2.5 vs DeepSeek V3.215Kimi K2.5 vs GPT-5.4 Pro14Kimi K2.5 vs Llama 3 8B Instruct11Kimi K2.5 vs Llama 3.2 1B11Kimi K2.5 vs Llama 2 70B Chat10Kimi K2.5 vs GPT-5.4-Cyber9Kimi K2.5 vs o3 Deep Research9Kimi K2.5 vs Qwen2.5-72B8Kimi K2.5 vs Together AI - Llama 3 8B Lite7Kimi K2.5 vs GPT-5.46Kimi K2.5 vs Mixtral 8x7B5Kimi K2.5 vs Qwen2-7B-Instruct4Kimi K2.5 vs Qwen3.5-122B-A10B4Kimi K2.5 vs Mixtral 8x22B Instruct v0.34Kimi K2.5 vs StepFun Step-24Kimi K2.5 vs Llama 3.2 1B Instruct3Kimi K2.5 vs Qwen3.5-9B3Kimi K2.5 vs Gemini 2.5 Pro Preview 05-062Kimi K2.5 vs Phi-4 Mini Flash Reasoning2Kimi K2.5 vs Qwen2.5-Max1Kimi K2.5 vs Qwen3-Max0

Frequently asked questions

What is the context window of Kimi K2.5?

Kimi K2.5 has a context window of 256k tokens.

How much does Kimi K2.5 cost?

Kimi K2.5 pricing ranges from $0.44/1M to $0.6/1M input tokens depending on the provider.

When was Kimi K2.5 released?

Kimi K2.5 was released on 2026-03-15.

Which providers offer Kimi K2.5?

Kimi K2.5 is available from 10 providers: Cloudflare Workers AI, Fireworks AI, OpenRouter, Together AI, NVIDIA NIM, AWS Bedrock, Replicate API, Microsoft Foundry, Vercel AI Gateway, Novita AI.

What benchmarks has Kimi K2.5 been tested on?

Kimi K2.5 has been evaluated on 17 benchmarks, including Google-Proof Q&A, MMLU PRO, BFCL, τ-bench, MultiChallenge.