Compare / Gemma 4 31B vs Qwen3 30B-A3B Thinking 2507

Gemma 4 31B vs Qwen3 30B-A3B Thinking 2507

Updated November 2026

Verdict

Gemma 4 31B scores higher on the Small LM Index: 70.1 vs 59.3. On the 6 benchmarks both labs report, Gemma 4 31B wins 6 and Qwen3 30B-A3B Thinking 2507 wins 0. Both need about 21.9 GB of memory at 4-bit quantization. Only Gemma 4 31B has a hosted API on OpenRouter ($0.16 per million tokens); the other has to be self-hosted.

Pick Gemma 4 31B if…

you want the strongest model overall
you mostly write code
you need math and reasoning
you want a hosted API rather than running it yourself
you need image input
you want the more widely used model (more quantized builds and community support)

Pick Qwen3 30B-A3B Thinking 2507 if…

you work with long documents

Specifications

	Gemma 4 31B	Qwen3 30B-A3B Thinking 2507
Small LM Index	70.1 (likely 69–72)	59.3 (likely 55–63)
Lab	Google · United States	Alibaba (Qwen) · China
Released	2026-04-02	2025-07-30
Parameters	30.7B (dense)	30.5B (3.3B active, MoE)
Memory at 4-bit	~21.9 GB	~20.4 GB
Context window	256K tokens	262K tokens
Input	text, image	text
Reasoning	Hybrid (thinking on/off)	Always thinks
License	Apache-2.0	Apache-2.0
API price / 1M tokens	$0.16 ($0.10 in, $0.34 out)	Self-host only
Hugging Face downloads (30 days)	8M	66.6K

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	Gemma 4 31B	Qwen3 30B-A3B Thinking 2507
MMLU-Pro Knowledge	85.2	80.9
GPQA Diamond Knowledge	84.3	73.4
Humanity's Last Exam Reasoning	19.5	9.8
AIME Math	89.2	85.0
LiveCodeBench Coding	80.0	66.0
SWE-bench Verified Coding	52.0^*	22.0^*
Terminal-Bench Agentic	42.9^*	—
τ²-Bench Agentic	76.9	—
IFBench Instruction following	—	51.5
MMMU-Pro Vision	76.9	—

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, Gemma 4 31B or Qwen3 30B-A3B Thinking 2507?

Gemma 4 31B scores higher on the Small LM Index: 70.1 vs 59.3. On the 6 benchmarks both labs report, Gemma 4 31B wins 6 and Qwen3 30B-A3B Thinking 2507 wins 0.

Which is cheaper to run?

Only Gemma 4 31B has a hosted API on OpenRouter ($0.16 per million tokens); the other has to be self-hosted. Both need about 21.9 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 Gemma 4 31B and Qwen3 30B-A3B Thinking 2507 commercially?

Both are released under the Apache-2.0 license. Apache 2.0 and MIT allow commercial use; other licenses can add conditions, so read the license before shipping a product.