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。
GLM-4.7-Flash
智谱AI
Gemma 4 26B A4B
DeepMind
Each axis is a category average, normalized to a 100-point radar.
Relative edge: Agent能力评测 +11.3 / Relative gap: 科学与综合推理 -11.7
Relative edge: 科学与综合推理 +11.7 / Relative gap: Agent能力评测 -11.3
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
GLM-4.7-Flash · 56.37
Best single
Gemma 4 26B A4B · GPQA Diamond 82.30
Modality coverage
Gemma 4 26B A4B · 2 modalities
Head to head
3
Benchmarks
1
Wins
2
Losses
+0.47
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 | GLM-4.7-Flash | Gemma 4 26B A4B |
|---|---|---|
HLE 综合评估 | 14.40Thinking Enabled | 17.20Thinking Enabled ｜ Tools |
GPQA Diamond 科学与综合推理 | 75.20Thinking Enabled | 82.30Thinking Enabled |
τ²-Bench Agent能力评测 | 79.50Thinking Enabled ｜ Tools | 68.20Thinking Enabled ｜ Tools |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | GLM-4.7-Flash智谱AI | Gemma 4 26B A4BDeepMind |
|---|---|---|
Core specsRelease | 2026-01-19 | 2026-04-02 |
Context length | 200K | 256K |
Parameters | 31B | 25.2B |
Active parameters | 3B | 3.8B |
Max output | 131072 | 32768 |
MoE | Yes | Yes |
Supported modes | 常规模式（Non-Thinking Mode）思考模式（Thinking Mode） | No mode data |
LicenseCode Open Source | Open Source · MIT License | Open Source · Apache 2.0 |
Weights Open Source | Open Source · MIT License | Open Source · Apache 2.0 |
Commercial use | 免费商用授权 | 免费商用授权 |
Local deploymentWeight size | 62.5GB | Not provided |
VRAM for weights | ≈ 62.5 GB (weights only, excludes KV cache) | Not provided |
Weights | Hugging Face | Hugging Face |
Source repo | GitHub | Not provided |
Modality supportText Input/Output | / | / |
Image Input/Output | Not provided | / |
ResourcesPaper / report | GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models (related technical report referenced on model card) | Gemma 4 Model Card |
DataLearner blog | Not provided | 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 |
