Model Comparison

Kimi K2-Instruct-0905 vs Qwen3 235B A22B

Kimi K2-Instruct-0905 has a slight edge in benchmark performance.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

10 benchmarks

Kimi K2-Instruct-0905 outperforms in 6 benchmarks (GPQA, MMLU, MMLU-Pro, MMLU-Redux, MultiPL-E, SuperGPQA), while Qwen3 235B A22B is better at 4 benchmarks (AIME 2024, AIME 2025, LiveBench, LiveCodeBench).

Kimi K2-Instruct-0905 has a slight edge in benchmark performance.

Mon May 25 2026 • llm-stats.com

Arena Performance

Human preference votes

Model Size

Parameter count comparison

765.0B diff

Kimi K2-Instruct-0905 has 765.0B more parameters than Qwen3 235B A22B, making it 325.5% larger.

Kimi K2-Instruct-0905

1.0Tparameters

Qwen3 235B A22B

235.0Bparameters

1000.0B

Kimi K2-Instruct-0905

235.0B

Qwen3 235B A22B

Context Window

Maximum input and output token capacity

Only Qwen3 235B A22B specifies input context (128,000 tokens). Only Qwen3 235B A22B specifies output context (128,000 tokens).

Kimi K2-Instruct-0905

Input- tokens

Output- tokens

Qwen3 235B A22B

Input128,000 tokens

Output128,000 tokens

Mon May 25 2026 • llm-stats.com

License

Usage and distribution terms

Kimi K2-Instruct-0905 is licensed under MIT, while Qwen3 235B A22B uses Apache 2.0.

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

Kimi K2-Instruct-0905

MIT

Open weights

Qwen3 235B A22B

Apache 2.0

Open weights

Release Timeline

When each model was launched

Kimi K2-Instruct-0905 was released on 2025-09-05, while Qwen3 235B A22B was released on 2025-04-29.

Kimi K2-Instruct-0905 is 4 months newer than Qwen3 235B A22B.

Kimi K2-Instruct-0905

Sep 5, 2025

8 months ago

4mo newer

Qwen3 235B A22B

Apr 29, 2025

1.1 years 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

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

Kimi K2-Instruct-0905

View details

Moonshot AI

Higher GPQA score (75.1% vs 47.5%)

Higher MMLU score (89.5% vs 87.8%)

Higher MMLU-Pro score (81.1% vs 68.2%)

Higher MMLU-Redux score (92.7% vs 87.4%)

Higher MultiPL-E score (85.7% vs 65.9%)

Higher SuperGPQA score (57.2% vs 44.1%)

Qwen3 235B A22B

View details

Alibaba Cloud / Qwen Team

Larger context window (128,000 tokens)

Higher AIME 2024 score (85.7% vs 69.6%)

Higher AIME 2025 score (81.5% vs 49.5%)

Higher LiveBench score (77.1% vs 76.4%)

Higher LiveCodeBench score (70.7% vs 53.7%)

Detailed Comparison

AI Model Comparison Table
Feature	Kimi K2-Instruct-0905	Qwen3 235B A22B

FAQ

Common questions about Kimi K2-Instruct-0905 vs Qwen3 235B A22B.

Which is better, Kimi K2-Instruct-0905 or Qwen3 235B A22B?

Kimi K2-Instruct-0905 has a slight edge in benchmark performance. Kimi K2-Instruct-0905 is made by Moonshot AI and Qwen3 235B A22B is made by Alibaba Cloud / Qwen Team. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.

How does Kimi K2-Instruct-0905 compare to Qwen3 235B A22B in benchmarks?

Kimi K2-Instruct-0905 scores MATH-500: 97.4%, MMLU-Redux: 92.7%, IFEval: 89.8%, AutoLogi: 89.5%, MMLU: 89.5%. Qwen3 235B A22B scores Arena Hard: 95.6%, GSM8k: 94.4%, BBH: 88.9%, MMLU: 87.8%, MMLU-Redux: 87.4%.

What are the context window sizes for Kimi K2-Instruct-0905 and Qwen3 235B A22B?

Kimi K2-Instruct-0905 supports an unknown number of tokens and Qwen3 235B A22B supports 128K 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-Instruct-0905 and Qwen3 235B A22B?

Key differences include licensing (MIT vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes Kimi K2-Instruct-0905 and Qwen3 235B A22B?

Kimi K2-Instruct-0905 is developed by Moonshot AI and Qwen3 235B A22B is developed by Alibaba Cloud / Qwen Team.