The AI arena is free today

Open Superagent

Kimi K2.5 vs Qwen3 VL 30B A3B Instruct

Kimi K2.5 leads the LLM Stats Score 38.9 to 16.5. Qwen3 VL 30B A3B Instruct is 4.6x cheaper per token.

Moonshot AI · Alibaba Cloud / Qwen Team · Updated for 2026

Which is better?

Kimi K2.5 leads the overall LLM Stats Score 38.9 to 16.5, ranking #60 overall.

In the 13 individual benchmarks reported for both models, Kimi K2.5 wins 13; this is a narrower head-to-head signal than the composite indexes.

On price, Qwen3 VL 30B A3B Instruct is roughly 4.6x cheaper per token on a blended 3:1 input/output basis, which adds up quickly at production volume.

Qwen3 VL 30B A3B Instruct also accepts a larger context window (262,144 input tokens), making it the stronger choice for long documents and large codebases.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose Kimi K2.5

  • overall performance matters — it scores 38.9 and ranks #60 on LLM Stats
  • your work emphasizes reasoning and agents — it leads those capability indexes
  • you value its reported benchmark strengths — it wins 13 of 13 exact shared results
  • you want the most recent training data — it shipped Jan 2026

Choose Qwen3 VL 30B A3B Instruct

  • cost matters — it's about 4.6x cheaper per token
  • you process long inputs — it offers a 262,144 token context window

At a glance

The differences that matter most.

Core performance indexes
38.9
#60
16.5
#221
38.8
#58
15.1
#224
18.5
#67
4.5
#153
Cost, coverage & limits
Benchmark wins
13 of 13
0 of 13
Input price
$0.60 / M
$0.15 / M
Output price
$3.00 / M
$0.60 / M
Context window
262,100
262,144

Capability indexes

Additional strengths measured across groups of related public benchmarks

3 shared
Index
Kimi K2.5
Qwen3 VL 30B A3B Instruct
36.6#36
16.5#202
27.4#37
11.3#108
29.4#29
13.8#94
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

40 reported for Kimi K2.5 · 50 for Qwen3 VL 30B A3B Instruct

13 shared

Kimi K2.5 outperforms in 13 benchmarks (AIME 2025, CharXiv-R, GPQA, InfoVQAtest, LiveCodeBench v6, LVBench, MathVision, MathVista-Mini, MMLU-Pro, MMMU-Pro, OCRBench, Video-MME, VideoMMMU), while Qwen3 VL 30B A3B Instruct is better at 0 benchmarks.

Kimi K2.5 significantly outperforms across most benchmarks.

Thu Sep 10 2026 • llm-stats.com

Human preference

Blind head-to-head votes and playground preference scores

Pricing Analysis

Price comparison per million tokens

Qwen3 VL 30B A3B Instruct costs less

For input processing, Kimi K2.5 ($0.60/1M tokens) is 4.0x more expensive than Qwen3 VL 30B A3B Instruct ($0.15/1M tokens).

For output processing, Kimi K2.5 ($3.00/1M tokens) is 5.0x more expensive than Qwen3 VL 30B A3B Instruct ($0.60/1M tokens).

In conclusion, Kimi K2.5 is more expensive than Qwen3 VL 30B A3B Instruct.*

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

Lowest available price from all providers
Thu Sep 10 2026 • llm-stats.com
Moonshot AI
Kimi K2.5
Input tokens$0.60
Output tokens$3.00
Best providerFireworks
Alibaba Cloud / Qwen Team
Qwen3 VL 30B A3B Instruct
Input tokens$0.15
Output tokens$0.60
Best providerDeepinfra
Notice missing or incorrect data?Start an Issue

Model Size

Parameter count comparison

969.0B diff

Kimi K2.5 has 969.0B more parameters than Qwen3 VL 30B A3B Instruct, making it 3125.8% larger.

Moonshot AI
Kimi K2.5
1.0Tparameters
Alibaba Cloud / Qwen Team
Qwen3 VL 30B A3B Instruct
31.0Bparameters
1000.0B
Kimi K2.5
31.0B
Qwen3 VL 30B A3B Instruct

Context Window

Maximum input and output token capacity

Qwen3 VL 30B A3B Instruct accepts 262,144 input tokens compared to Kimi K2.5's 262,100 tokens. Qwen3 VL 30B A3B Instruct 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
Alibaba Cloud / Qwen Team
Qwen3 VL 30B A3B Instruct
Input262,144 tokens
Output262,144 tokens
Thu Sep 10 2026 • llm-stats.com

Input capabilities

Documented input modalities across available providers

Both Kimi K2.5 and Qwen3 VL 30B A3B Instruct support multimodal inputs.

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

Kimi K2.5

Text
Images
Audio
Video

Qwen3 VL 30B A3B Instruct

Text
Images
Audio
Video

License

Usage and distribution terms

Kimi K2.5 is licensed under MIT, while Qwen3 VL 30B A3B Instruct 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

Qwen3 VL 30B A3B Instruct

Apache 2.0

Open weights

Release Timeline

When each model was launched

Kimi K2.5 was released on 2026-01-27, while Qwen3 VL 30B A3B Instruct was released on 2025-09-22.

Kimi K2.5 is 4 months newer than Qwen3 VL 30B A3B Instruct.

Kimi K2.5

Jan 27, 2026

7 months ago

4mo newer
Qwen3 VL 30B A3B Instruct

Sep 22, 2025

11 months ago

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. Qwen3 VL 30B A3B Instruct is available from DeepInfra, 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

Qwen3 VL 30B A3B Instruct

deepinfra logo
Deepinfra
Input Price:Input: $0.15/1MOutput Price:Output: $0.60/1M
novita logo
Novita
Input Price:Input: $0.20/1MOutput Price:Output: $0.70/1M
* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Judge for yourself.

Run your own prompts against Kimi K2.5 and Qwen3 VL 30B A3B Instruct side-by-side, then vote on the output you prefer.

Kimi K2.5
✓ Preferred
Qwen3 VL 30B A3B Instruct
Open in Playground

FAQ

Common questions about Kimi K2.5 vs Qwen3 VL 30B A3B Instruct.

Which is better, Kimi K2.5 or Qwen3 VL 30B A3B Instruct?

Kimi K2.5 leads the LLM Stats Score 38.9 to 16.5. Kimi K2.5 is made by Moonshot AI and Qwen3 VL 30B A3B Instruct is made by Alibaba Cloud / Qwen Team. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does Kimi K2.5 compare to Qwen3 VL 30B A3B Instruct in benchmarks?

Kimi K2.5 scores AIME 2025: 96.1%, HMMT 2025: 95.4%, InfoVQAtest: 92.6%, OCRBench: 92.3%, MathVista-Mini: 90.1%. Qwen3 VL 30B A3B Instruct scores DocVQAtest: 95.0%, ScreenSpot: 94.7%, OCRBench: 90.3%, MMLU-Redux: 88.4%, MMBench-V1.1: 87.0%.

Is Kimi K2.5 cheaper than Qwen3 VL 30B A3B Instruct?

Qwen3 VL 30B A3B Instruct is 4.0x cheaper for input tokens. Kimi K2.5 costs $0.60/M input and $3.00/M output via fireworks. Qwen3 VL 30B A3B Instruct costs $0.15/M input and $0.60/M output via deepinfra.

What are the context window sizes for Kimi K2.5 and Qwen3 VL 30B A3B Instruct?

Kimi K2.5 supports 262K tokens and Qwen3 VL 30B A3B Instruct supports 262K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Kimi K2.5 and Qwen3 VL 30B A3B Instruct?

Key differences include LLM Stats Score (38.9 vs 16.5), context window (262K vs 262K), input pricing ($0.60 vs $0.15/M), licensing (MIT vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes Kimi K2.5 and Qwen3 VL 30B A3B Instruct?

Kimi K2.5 is developed by Moonshot AI and Qwen3 VL 30B A3B Instruct is developed by Alibaba Cloud / Qwen Team.