Model Comparison

Kimi K2.5 vs Kimi K2.6

Kimi K2.6 significantly outperforms across most benchmarks. Kimi K2.5 is 1.4x cheaper per token.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

15 benchmarks

Kimi K2.5 outperforms in 1 benchmarks (Humanity's Last Exam), while Kimi K2.6 is better at 14 benchmarks (BrowseComp, CharXiv-R, DeepSearchQA, GPQA, IMO-AnswerBench, LiveCodeBench v6, MathVision, MMMU-Pro, SciCode, SWE-bench Multilingual, SWE-Bench Pro, SWE-Bench Verified, Terminal-Bench 2.0, WideSearch).

Kimi K2.6 significantly outperforms across most benchmarks.

Mon Apr 20 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

Kimi K2.5 costs less

For input processing, Kimi K2.5 ($0.60/1M tokens) is 1.6x cheaper than Kimi K2.6 ($0.95/1M tokens).

For output processing, Kimi K2.5 ($3.00/1M tokens) is 1.3x cheaper than Kimi K2.6 ($4.00/1M tokens).

In conclusion, Kimi K2.6 is more expensive than Kimi K2.5.*

* Using a 3:1 ratio of input to output tokens

Lowest available price from all providers
Mon Apr 20 2026 • llm-stats.com
Moonshot AI
Kimi K2.5
Input tokens$0.60
Output tokens$3.00
Best providerFireworks
Moonshot AI
Kimi K2.6
Input tokens$0.95
Output tokens$4.00
Best providerUnknown Organization
Notice missing or incorrect data?Start an Issue→

Model Size

Parameter count comparison

0.0M diff

Kimi K2.6 has 0.0B more parameters than Kimi K2.5, making it 0.0% larger.

Moonshot AI
Kimi K2.5
1000.0Bparameters
Moonshot AI
Kimi K2.6
1000.0Bparameters
1000.0B
Kimi K2.5
1000.0B
Kimi K2.6

Context Window

Maximum input and output token capacity

Kimi K2.6 accepts 262,144 input tokens compared to Kimi K2.5's 262,100 tokens. Kimi K2.6 can generate longer responses up to 262,144 tokens, while Kimi K2.5 is limited to 262,100 tokens.

Moonshot AI
Kimi K2.5
Input262,100 tokens
Output262,100 tokens
Moonshot AI
Kimi K2.6
Input262,144 tokens
Output262,144 tokens
Mon Apr 20 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Both Kimi K2.5 and Kimi K2.6 support multimodal inputs.

They are both capable of processing various types of data, offering versatility in application.

Kimi K2.5

Text
Images
Audio
Video

Kimi K2.6

Text
Images
Audio
Video

License

Usage and distribution terms

Kimi K2.5 is licensed under MIT, while Kimi K2.6 uses Modified MIT License.

License differences may affect how you can use these models in commercial or open-source projects.

Kimi K2.5

MIT

Open weights

Kimi K2.6

Modified MIT License

Open weights

Release Timeline

When each model was launched

Kimi K2.5 was released on 2026-01-27, while Kimi K2.6 was released on 2026-04-20.

Kimi K2.6 is 3 months newer than Kimi K2.5.

Kimi K2.5

Jan 27, 2026

2 months ago

Kimi K2.6

Apr 20, 2026

0 days ago

2mo newer

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Provider Availability

Kimi K2.5 is available from Fireworks, Moonshot AI. Kimi K2.6 is available from Moonshot AI, Novita.

Kimi K2.5

fireworks logo
Fireworks
Input Price:Input: $0.60/1MOutput Price:Output: $3.00/1M
moonshot logo
Unknown Organization
Input Price:Input: $0.60/1MOutput Price:Output: $3.00/1M

Kimi K2.6

moonshot logo
Unknown Organization
Input Price:Input: $0.95/1MOutput Price:Output: $4.00/1M
novita logo
Novita
Input Price:Input: $0.95/1MOutput Price:Output: $4.00/1M
* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

Moonshot AI
Less expensive input tokens
Less expensive output tokens
Higher Humanity's Last Exam score (50.2% vs 36.4%)
Moonshot AI
Larger context window (262,144 tokens)
Higher BrowseComp score (86.3% vs 74.9%)
Higher CharXiv-R score (86.7% vs 77.5%)
Higher DeepSearchQA score (83.0% vs 77.1%)
Higher GPQA score (90.5% vs 87.6%)
Higher IMO-AnswerBench score (86.0% vs 81.8%)
Higher LiveCodeBench v6 score (89.6% vs 85.0%)
Higher MathVision score (93.2% vs 84.2%)
Higher MMMU-Pro score (80.1% vs 78.5%)
Higher SciCode score (52.2% vs 48.7%)
Higher SWE-bench Multilingual score (76.7% vs 73.0%)
Higher SWE-Bench Pro score (58.6% vs 50.7%)
Higher SWE-Bench Verified score (80.2% vs 76.8%)
Higher Terminal-Bench 2.0 score (66.7% vs 50.8%)
Higher WideSearch score (80.8% vs 79.0%)

Detailed Comparison

AI Model Comparison Table
Feature
Moonshot AI
Kimi K2.5
Moonshot AI
Kimi K2.6

FAQ

Common questions about Kimi K2.5 vs Kimi K2.6

Kimi K2.6 significantly outperforms across most benchmarks. Kimi K2.5 is made by Moonshot AI and Kimi K2.6 is made by Moonshot AI. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.
Kimi K2.5 scores AIME 2025: 96.1%, HMMT 2025: 95.4%, InfoVQAtest: 92.6%, OCRBench: 92.3%, MathVista-Mini: 90.1%. Kimi K2.6 scores V*: 96.9%, AIME 2026: 96.4%, MathVision: 93.2%, HMMT Feb 26: 92.7%, GPQA: 90.5%.
Kimi K2.5 is 1.6x cheaper for input tokens. Kimi K2.5 costs $0.60/M input and $3.00/M output via fireworks. Kimi K2.6 costs $0.95/M input and $4.00/M output via moonshot.
Kimi K2.5 supports 262K tokens and Kimi K2.6 supports 262K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.
Key differences include context window (262K vs 262K), input pricing ($0.60 vs $0.95/M), licensing (MIT vs Modified MIT License). See the full comparison above for benchmark-by-benchmark results.