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。
Qwen3.8-27B
阿里巴巴
Each axis is a category average, normalized to a 100-point radar.
Relative edge: none clear / Relative gap: AI Agent - 工具使用 -19.9
Relative edge: AI Agent - 工具使用 +19.9 / Relative gap: none clear
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
Qwen3.8-27B · 74.33
Best single
Qwen3.8-27B · OmniDocBench 91.10
Modality coverage
Qwen3.8-27B · 3 modalities
Head to head
7
Benchmarks
0
Wins
7
Losses
-13.06
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.
7 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Muse Glimmer-30B | Qwen3.8-27B |
|---|---|---|
HLE 综合评估 | 22.00Thinking Level · High | 30.80Thinking Enabled |
GPQA Diamond 科学与综合推理 | 83.50Thinking Level · High | 89.20Thinking Enabled |
SWE-Bench Pro - Public 编程与软件工程 | 51.20Thinking Level · High ｜ Tools | 61.70Thinking Enabled ｜ Tools |
OSWorld-Verified AI Agent - 工具使用 | 65.90Thinking Level · High ｜ Tools | 84.30Thinking Enabled ｜ Tools |
Terminal-Bench 2.1 AI Agent - 工具使用 | 51.70Thinking Level · High ｜ Tools | 73.00Thinking Enabled ｜ Tools |
CharXiv RQ 多模态理解 | 78.80Thinking Level · High | 90.20Thinking Enabled ｜ Tools |
OmniDocBench 多模态理解 | 75.80Thinking Level · High | 91.10Thinking Enabled |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | Muse Glimmer-30BFacebook AI研究实验室 | Qwen3.8-27B阿里巴巴 |
|---|---|---|
Core specsRelease | 2026-08-10 | 2026-08-14 |
Context length | 128K+ | 256K |
Parameters | 29.6B | 27B |
Active parameters | 29.6B | 27B |
Max output | Not provided | 131072 |
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 | 约 55.5 GiB（BF16）；4-bit 量化版本低于 20 GB | 约 55.6 GB（BF16，18 个 safetensors 分片） |
VRAM for weights | ≈ 55.5 GB (weights only, excludes KV cache) | ≈ 55.6 GB (weights only, excludes KV cache) |
Weights | Hugging Face | Hugging Face |
Modality supportText Input/Output | / | / |
Image Input/Output | / | / |
Video Input/Output | Not provided | / |
ResourcesPaper / report | Introducing Muse Glimmer: An Open Agentic Model That Runs on Your Device | 阿里Qwen3.8-Max模型发布：2.4万亿参数，瞄准复杂长程任务 |
