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。
Haiku 4.5
Anthropic
Each axis is a category average, normalized to a 100-point radar.
Relative edge: OpenClaw智能体能力综合测评 +17.0 / Relative gap: 编程与软件工程 -17.2
Relative edge: 编程与软件工程 +17.2 / Relative gap: OpenClaw智能体能力综合测评 -17.0
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
Qwen3.6-27B · 68.82
Best single
Haiku 4.5 · Claw Bench 89.40
Modality coverage
Qwen3.6-27B · 3 modalities
Head to head
8
Benchmarks
1
Wins
7
Losses
-7.76
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.
8 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Haiku 4.5 | Qwen3.6-27B |
|---|---|---|
GPQA Diamond 综合评估 | 73.30Extended Thinking | 87.80Thinking Enabled |
HLE 综合评估 | 9.70Extended Thinking | 24.00Thinking Enabled |
LiveBench 综合评估 | 61.3264K | 65.56Standard Mode |
MMLU Pro 综合评估 | 80.00Extended Thinking | 86.20Thinking Enabled |
LiveCodeBench 编程与软件工程 | 62.00Extended Thinking | 83.90Thinking Enabled |
SWE-Bench Pro - Public 编程与软件工程 | 39.45Extended Thinking ｜ Tools | 53.50Thinking Enabled ｜ Tools |
SWE-bench Verified 编程与软件工程 | 73.30128K ｜ Tools | 77.20Thinking Enabled ｜ Tools |
Claw Bench OpenClaw智能体能力综合测评 | 89.40Thinking Enabled ｜ Tools | 72.40Thinking Enabled ｜ Tools |
Side-by-side input/output token pricing
Licensing, MoE architecture, and multi-modality support.
| Features & specs | Haiku 4.5Anthropic | Qwen3.6-27B阿里巴巴 |
|---|---|---|
Core specsRelease | 2025-10-15 | 2026-04-22 |
Context length | 200K | 128K |
Parameters | — | 270 |
Active parameters | Not provided | 270 |
Max output | 65536 | 16384 |
MoE | No | No |
Supported modes | 常规模式（Non-Thinking Mode）思考模式（Thinking Mode） | No mode data |
LicenseCode Open Source | Not provided | Not provided |
Weights Open Source | Not provided | Not provided |
Commercial use | 不开源 | 免费商用授权 |
Modality supportText Input/Output | / | / |
Image Input/Output | / | / |
Video Input/Output | Not provided | / |
ResourcesPaper / report | Introducing Claude Haiku 4.5 | Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model |
DataLearner blog | Anthropic发布Claude Haiku 4.5：一个不可忽视的低成本模型，1/3的价格，1.5倍的速度，但是有Claude Sonnet 4的水平！ | 阿里正式开源Qwen3.6-27B：代码智能体能力上超越全面超越前代旗舰版本之 Qwen3.5-397B-A17B |
Qwen3.6-27B
阿里巴巴
