See key specs and per-benchmark scores for each model/mode. Scroll horizontally for all columns. 当前对比 2 个模型的评测数据与核心参数。

Gemma 4 31B
DeepMind
Each axis is a category average, normalized to a 100-point radar.
Relative edge: 综合评估 +1.0 / Relative gap: 数学推理 -5.5
Relative edge: 数学推理 +5.5 / Relative gap: 综合评估 -1.0
Method: for each model and benchmark, the chart first averages all scores in the current mode scope instead of taking the best score, then averages those benchmark scores within each category. Only benchmarks with at least two selected models scored are included; missing values are not counted as zero.
Best overall
Muse Glimmer-30B · 66.73
Best single
Muse Glimmer-30B · AIME 2026 94.70
Modality coverage
Gemma 4 31B · 2 modalities
Head to head
3
Benchmarks
2
Wins
1
Losses
-0.07
Average diff
Compare benchmark results across thinking modes and tool usage.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
Complete scores for each model/mode across selected benchmarks.
3 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Gemma 4 31B | Muse Glimmer-30B |
|---|---|---|
HLE 综合评估 | 26.50Thinking Enabled | Tools | 22.00Thinking Level · High |
GPQA Diamond 科学与综合推理 | 84.30Thinking Enabled | 83.50Thinking Level · High |
AIME 2026 数学推理 | 89.20Thinking Enabled | 94.70Thinking Level · High |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | Gemma 4 31BDeepMind | Muse Glimmer-30BFacebook AI研究实验室 |
|---|---|---|
Core specsRelease | 2026-04-02 | 2026-08-10 |
Context length | 256K | 128K+ |
Parameters | 30.7B | 29.6B |
Active parameters | Not provided | 29.6B |
Max output | 32768 | Not provided |
MoE | No | No |
LicenseCode Open Source | Open Source · Apache 2.0 | Open Source · Apache 2.0 |
Weights Open Source | Open Source · Apache 2.0 | Open Source · Apache 2.0 |
Commercial use | 免费商用授权 | 免费商用授权 |
Local deploymentWeight size | Not provided | 约 55.5 GiB(BF16);4-bit 量化版本低于 20 GB |
VRAM for weights | Not provided | ≈ 55.5 GB (weights only, excludes KV cache) |
Weights | Hugging Face | Hugging Face |
Modality supportText Input/Output | / | / |
Image Input/Output | / | / |
ResourcesPaper / report | Gemma 4 Model Card | Introducing Muse Glimmer: An Open Agentic Model That Runs on Your Device |
DataLearner blog | Google Gemma 4 正式开源:Apache 2.0 协议、手机端可运行、原生支持多模态和 Agent 工作流 | Not provided |