Model Comparison

Kimi K2.5 vs Gemma 4 31B

Kimi K2.5 significantly outperforms across most benchmarks. Gemma 4 31B is 5.9x cheaper per token.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

6 benchmarks

Kimi K2.5 outperforms in 5 benchmarks (GPQA, Humanity's Last Exam, LiveCodeBench v6, MMLU-Pro, MMMU-Pro), while Gemma 4 31B is better at 1 benchmark (MathVision).

Kimi K2.5 significantly outperforms across most benchmarks.

Sat Apr 18 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

Gemma 4 31B costs less

For input processing, Kimi K2.5 ($0.60/1M tokens) is 4.3x more expensive than Gemma 4 31B ($0.14/1M tokens).

For output processing, Kimi K2.5 ($3.00/1M tokens) is 7.5x more expensive than Gemma 4 31B ($0.40/1M tokens).

In conclusion, Kimi K2.5 is more expensive than Gemma 4 31B.*

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

Lowest available price from all providers
Sat Apr 18 2026 • llm-stats.com
Moonshot AI
Kimi K2.5
Input tokens$0.60
Output tokens$3.00
Best providerFireworks
Google
Gemma 4 31B
Input tokens$0.14
Output tokens$0.40
Best providerNovita
Notice missing or incorrect data?Start an Issue→

Model Size

Parameter count comparison

969.3B diff

Kimi K2.5 has 969.3B more parameters than Gemma 4 31B, making it 3157.3% larger.

Moonshot AI
Kimi K2.5
1000.0Bparameters
Google
Gemma 4 31B
30.7Bparameters
1000.0B
Kimi K2.5
30.7B
Gemma 4 31B

Context Window

Maximum input and output token capacity

Gemma 4 31B accepts 262,144 input tokens compared to Kimi K2.5's 262,100 tokens. Kimi K2.5 can generate longer responses up to 262,100 tokens, while Gemma 4 31B is limited to 131,072 tokens.

Moonshot AI
Kimi K2.5
Input262,100 tokens
Output262,100 tokens
Google
Gemma 4 31B
Input262,144 tokens
Output131,072 tokens
Sat Apr 18 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Both Kimi K2.5 and Gemma 4 31B support multimodal inputs.

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

Kimi K2.5

Text
Images
Audio
Video

Gemma 4 31B

Text
Images
Audio
Video

License

Usage and distribution terms

Kimi K2.5 is licensed under MIT, while Gemma 4 31B uses Apache 2.0.

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

Kimi K2.5

MIT

Open weights

Gemma 4 31B

Apache 2.0

Open weights

Release Timeline

When each model was launched

Kimi K2.5 was released on 2026-01-27, while Gemma 4 31B was released on 2026-04-02.

Gemma 4 31B is 2 months newer than Kimi K2.5.

Kimi K2.5

Jan 27, 2026

2 months ago

Gemma 4 31B

Apr 2, 2026

2 weeks ago

2mo newer

Knowledge Cutoff

When training data ends

Gemma 4 31B has a documented knowledge cutoff of 2025-01-01, while Kimi K2.5's cutoff date is not specified.

We can confirm Gemma 4 31B's training data extends to 2025-01-01, but cannot make a direct comparison without Kimi K2.5's cutoff date.

Kimi K2.5

—

Gemma 4 31B

Jan 2025

Provider Availability

Kimi K2.5 is available from Fireworks, Moonshot AI. Gemma 4 31B is available from 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

Gemma 4 31B

novita logo
Novita
Input Price:Input: $0.14/1MOutput Price:Output: $0.40/1M
* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

Moonshot AI
Higher GPQA score (87.6% vs 84.3%)
Higher Humanity's Last Exam score (50.2% vs 26.5%)
Higher LiveCodeBench v6 score (85.0% vs 80.0%)
Higher MMLU-Pro score (87.1% vs 85.2%)
Higher MMMU-Pro score (78.5% vs 76.9%)
Google
Larger context window (262,144 tokens)
Less expensive input tokens
Less expensive output tokens
Higher MathVision score (85.6% vs 84.2%)

Detailed Comparison

AI Model Comparison Table
Feature
Moonshot AI
Kimi K2.5
Google
Gemma 4 31B

FAQ

Common questions about Kimi K2.5 vs Gemma 4 31B

Kimi K2.5 significantly outperforms across most benchmarks. Kimi K2.5 is made by Moonshot AI and Gemma 4 31B is made by Google. 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%. Gemma 4 31B scores AIME 2026: 89.2%, MMMLU: 88.4%, t2-bench: 86.4%, MathVision: 85.6%, MMLU-Pro: 85.2%.
Gemma 4 31B is 4.3x cheaper for input tokens. Kimi K2.5 costs $0.60/M input and $3.00/M output via fireworks. Gemma 4 31B costs $0.14/M input and $0.40/M output via novita.
Kimi K2.5 supports 262K tokens and Gemma 4 31B 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 $0.14/M), licensing (MIT vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.
Kimi K2.5 is developed by Moonshot AI and Gemma 4 31B is developed by Google.