Compare / Qwen3 30B-A3B Thinking 2507 vs North Mini Code
Qwen3 30B-A3B Thinking 2507 vs North Mini Code
Updated November 2026
Verdict
Qwen3 30B-A3B Thinking 2507 and North Mini Code are roughly tied on the Small LM Index (65.4 vs 64.7). On the 6 benchmarks both labs report, Qwen3 30B-A3B Thinking 2507 wins 1 and North Mini Code wins 5. Both need about 21.4 GB of memory at 4-bit quantization. Neither model has a hosted API on OpenRouter, so you would run both yourself.
Pick Qwen3 30B-A3B Thinking 2507 if…
- you work with long documents
- you need a permissive license (Apache-2.0)
- you want the more widely used model (more quantized builds and community support)
Pick North Mini Code if…
- you mostly write code
- you need precise instruction following
- you build agents that call tools
Specifications
| Qwen3 30B-A3B Thinking 2507 | North Mini Code | |
|---|---|---|
| Small LM Index | 65.4 (likely 63–67) | 64.7 (likely 55–72) |
| Lab | Alibaba (Qwen) · China | Cohere · Other |
| Released | 2025-07-30 | 2026-06-17 |
| Parameters | 30.5B (3.3B active, MoE) | 30B (3B active, MoE) |
| Memory at 4-bit | ~20.4 GB | ~21.4 GB |
| Context window | 262K tokens | 256K tokens |
| Input | text | text |
| Reasoning | Always thinks | No thinking |
| License | Apache-2.0 | Unverified |
| API price / 1M tokens | Self-host only | Self-host only |
| Hugging Face downloads (30 days) | 68K | — |
Benchmarks
Higher is better. Bold marks the winner where both labs report a score. ≈ marks the index model's estimate for a score that wasn't reported, ✓ an independent measurement and * a score measured by another lab.
| Benchmark | Qwen3 30B-A3B Thinking 2507 | North Mini Code |
|---|---|---|
| MMLU-Pro Knowledge | 80.5✓ | ≈ 82 |
| GPQA Diamond Knowledge | 70.7✓ | 75.7✓ |
| Humanity's Last Exam Reasoning | 10.3✓ | 11.1✓ |
| AIME Math | 90.7✓ | ≈ 89 |
| LiveCodeBench Coding | 70.7✓ | ≈ 75 |
| SWE-bench Verified Coding | 22.0* | ≈ 56 |
| τ²-Bench Agentic | 28.1✓ | 37.4✓ |
| IFBench Instruction following | 50.7✓ | 57.6✓ |
| MATH-500 Math | 97.6✓ | — |
| AA-LCR Long context | 61.3✓ | 37.3✓ |
| SciCode Coding | 33.0✓ | 38.8✓ |
Scores are mostly reported by each lab with its own settings, so small gaps are noise. See the methodology. Open in the interactive comparison to add more models.
Questions
Which is better, Qwen3 30B-A3B Thinking 2507 or North Mini Code?
Qwen3 30B-A3B Thinking 2507 and North Mini Code are roughly tied on the Small LM Index (65.4 vs 64.7). On the 6 benchmarks both labs report, Qwen3 30B-A3B Thinking 2507 wins 1 and North Mini Code wins 5.
Which is cheaper to run?
Neither model has a hosted API on OpenRouter, so you would run both yourself. Both need about 21.4 GB of memory at 4-bit quantization.
Which has the longer context window?
Qwen3 30B-A3B Thinking 2507 has the longer context window: 262K tokens against 256K.
Can I use Qwen3 30B-A3B Thinking 2507 and North Mini Code commercially?
Qwen3 30B-A3B Thinking 2507 is released under Apache-2.0 and North Mini Code under Unverified. Apache 2.0 and MIT allow commercial use; other licenses can add conditions, so read the license before shipping a product.
Related comparisons
- Muse Glimmer 30B vs Qwen3 30B-A3B Thinking 2507
- Gemma 4 31B vs Qwen3 30B-A3B Thinking 2507
- GLM-4.7-Flash vs Qwen3 30B-A3B Thinking 2507
- Granite 4.2 30B vs Qwen3 30B-A3B Thinking 2507
- Qwen3 30B-A3B Thinking 2507 vs Qwen3-Coder 30B-A3B
- Kanana-2 30B-A3B Thinking vs Qwen3 30B-A3B Thinking 2507
- Muse Glimmer 30B vs North Mini Code
- Granite 4.2 30B vs North Mini Code
- North Mini Code vs Qwen3-Coder 30B-A3B