The AI arena is free today

Open Superagent

K-EXAONE-236B-A23B vs Qwen3-235B-A22B-Thinking-2507

K-EXAONE-236B-A23B shows notably better performance in the majority of benchmarks. K-EXAONE-236B-A23B is 1.4x cheaper per token.

LG AI Research · Alibaba Cloud / Qwen Team · Updated for 2026

Which is better?

K-EXAONE-236B-A23B outperforms in 2 benchmarks (AIME 2025, LiveCodeBench v6), while Qwen3-235B-A22B-Thinking-2507 is better at 1 benchmark (MMLU-Pro). K-EXAONE-236B-A23B shows notably better performance in the majority of benchmarks.

On price, K-EXAONE-236B-A23B is roughly 1.4x cheaper per token on a blended 3:1 input/output basis, which adds up quickly at production volume.

Qwen3-235B-A22B-Thinking-2507 also accepts a larger context window (262,144 input tokens), making it the stronger choice for long documents and large codebases.

Based on current benchmark, pricing, and model metadata for 2026.

Choose K-EXAONE-236B-A23B

  • you want the strongest raw capability — it leads on 2 of 3 shared benchmarks
  • cost matters — it's about 1.4x cheaper per token
  • you want the most recent training data — it shipped Dec 2025

Choose Qwen3-235B-A22B-Thinking-2507

  • you process long inputs — it offers a 262,144 token context window
  • you need open weights you can self-host or fine-tune

At a glance

The differences that matter most.

Benchmark wins
2 of 3
1 of 3
Input price
$0.60 / M
$0.30 / M
Output price
$1.00 / M
$3.00 / M
Context window
32,768
262,144
Released
Dec 2025
Jul 2025
License
Proprietary
Apache 2.0

Performance Benchmarks

Comparative analysis across standard metrics

3 benchmarks

K-EXAONE-236B-A23B outperforms in 2 benchmarks (AIME 2025, LiveCodeBench v6), while Qwen3-235B-A22B-Thinking-2507 is better at 1 benchmark (MMLU-Pro).

K-EXAONE-236B-A23B shows notably better performance in the majority of benchmarks.

Mon Aug 24 2026 • llm-stats.com

Arena Performance

Playground indexes and blind preference scores

Pricing Analysis

Price comparison per million tokens

K-EXAONE-236B-A23B costs less

For input processing, K-EXAONE-236B-A23B ($0.60/1M tokens) is 2.0x more expensive than Qwen3-235B-A22B-Thinking-2507 ($0.30/1M tokens).

For output processing, K-EXAONE-236B-A23B ($1.00/1M tokens) is 3.0x cheaper than Qwen3-235B-A22B-Thinking-2507 ($3.00/1M tokens).

In conclusion, Qwen3-235B-A22B-Thinking-2507 is more expensive than K-EXAONE-236B-A23B.*

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

Lowest available price from all providers
Mon Aug 24 2026 • llm-stats.com
LG AI Research
K-EXAONE-236B-A23B
Input tokens$0.60
Output tokens$1.00
Best providerFriendliAI
Alibaba Cloud / Qwen Team
Qwen3-235B-A22B-Thinking-2507
Input tokens$0.30
Output tokens$3.00
Best providerFireworks
Notice missing or incorrect data?Start an Issue

Model Size

Parameter count comparison

1.0B diff

K-EXAONE-236B-A23B has 1.0B more parameters than Qwen3-235B-A22B-Thinking-2507, making it 0.4% larger.

LG AI Research
K-EXAONE-236B-A23B
236.0Bparameters
Alibaba Cloud / Qwen Team
Qwen3-235B-A22B-Thinking-2507
235.0Bparameters
236.0B
K-EXAONE-236B-A23B
235.0B
Qwen3-235B-A22B-Thinking-2507

Context Window

Maximum input and output token capacity

Qwen3-235B-A22B-Thinking-2507 accepts 262,144 input tokens compared to K-EXAONE-236B-A23B's 32,768 tokens. Qwen3-235B-A22B-Thinking-2507 can generate longer responses up to 131,072 tokens, while K-EXAONE-236B-A23B is limited to 32,768 tokens.

LG AI Research
K-EXAONE-236B-A23B
Input32,768 tokens
Output32,768 tokens
Alibaba Cloud / Qwen Team
Qwen3-235B-A22B-Thinking-2507
Input262,144 tokens
Output131,072 tokens
Mon Aug 24 2026 • llm-stats.com

License

Usage and distribution terms

K-EXAONE-236B-A23B is licensed under a proprietary license, while Qwen3-235B-A22B-Thinking-2507 uses Apache 2.0.

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

K-EXAONE-236B-A23B

Proprietary

Closed source

Qwen3-235B-A22B-Thinking-2507

Apache 2.0

Open weights

Release Timeline

When each model was launched

K-EXAONE-236B-A23B was released on 2025-12-31, while Qwen3-235B-A22B-Thinking-2507 was released on 2025-07-25.

K-EXAONE-236B-A23B is 5 months newer than Qwen3-235B-A22B-Thinking-2507.

K-EXAONE-236B-A23B

Dec 31, 2025

7 months ago

5mo newer
Qwen3-235B-A22B-Thinking-2507

Jul 25, 2025

1.1 years ago

Knowledge Cutoff

When training data ends

K-EXAONE-236B-A23B has a documented knowledge cutoff of 2025-10-01, while Qwen3-235B-A22B-Thinking-2507's cutoff date is not specified.

We can confirm K-EXAONE-236B-A23B's training data extends to 2025-10-01, but cannot make a direct comparison without Qwen3-235B-A22B-Thinking-2507's cutoff date.

K-EXAONE-236B-A23B

Oct 2025

Qwen3-235B-A22B-Thinking-2507

Provider Availability

K-EXAONE-236B-A23B is available from FriendliAI. Qwen3-235B-A22B-Thinking-2507 is available from Fireworks, Novita.

K-EXAONE-236B-A23B

friendli logo
FriendliAI
Input Price:Input: $0.60/1MOutput Price:Output: $1.00/1M

Qwen3-235B-A22B-Thinking-2507

fireworks logo
Fireworks
Input Price:Input: $0.30/1MOutput Price:Output: $3.00/1M
novita logo
Novita
Input Price:Input: $0.30/1MOutput Price:Output: $3.00/1M
* 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 K-EXAONE-236B-A23B and Qwen3-235B-A22B-Thinking-2507 side-by-side, then vote on the output you prefer.

K-EXAONE-236B-A23B
✓ Preferred
Qwen3-235B-A22B-Thinking-2507
Open in Playground

FAQ

Common questions about K-EXAONE-236B-A23B vs Qwen3-235B-A22B-Thinking-2507.

Which is better, K-EXAONE-236B-A23B or Qwen3-235B-A22B-Thinking-2507?

K-EXAONE-236B-A23B shows notably better performance in the majority of benchmarks. K-EXAONE-236B-A23B is made by LG AI Research and Qwen3-235B-A22B-Thinking-2507 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 K-EXAONE-236B-A23B compare to Qwen3-235B-A22B-Thinking-2507 in benchmarks?

K-EXAONE-236B-A23B scores AIME 2025: 92.8%, MMMLU: 85.7%, MMLU-Pro: 83.8%, LiveCodeBench v6: 80.7%, t2-bench: 73.2%. Qwen3-235B-A22B-Thinking-2507 scores MMLU-Redux: 93.8%, AIME 2025: 92.3%, WritingBench: 88.3%, IFEval: 87.8%, Creative Writing v3: 86.1%.

Is K-EXAONE-236B-A23B cheaper than Qwen3-235B-A22B-Thinking-2507?

Qwen3-235B-A22B-Thinking-2507 is 2.0x cheaper for input tokens. K-EXAONE-236B-A23B costs $0.60/M input and $1.00/M output via friendli. Qwen3-235B-A22B-Thinking-2507 costs $0.30/M input and $3.00/M output via fireworks.

What are the context window sizes for K-EXAONE-236B-A23B and Qwen3-235B-A22B-Thinking-2507?

K-EXAONE-236B-A23B supports 33K tokens and Qwen3-235B-A22B-Thinking-2507 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 K-EXAONE-236B-A23B and Qwen3-235B-A22B-Thinking-2507?

Key differences include context window (33K vs 262K), input pricing ($0.60 vs $0.30/M), licensing (Proprietary vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes K-EXAONE-236B-A23B and Qwen3-235B-A22B-Thinking-2507?

K-EXAONE-236B-A23B is developed by LG AI Research and Qwen3-235B-A22B-Thinking-2507 is developed by Alibaba Cloud / Qwen Team.