← All Comparisons
Claude Haiku 4.5 vs GLM-4.7
A detailed comparison of Claude Haiku 4.5 (Anthropic) and GLM-4.7 (Zhipu AI) across pricing, performance, and features.
Pricing Comparison
| Metric | Claude Haiku 4.5 | GLM-4.7 | Difference |
|---|---|---|---|
| Input / 1M tokens | $0.80 | $0.60 | -25% |
| Output / 1M tokens | $4.00 | $2.20 | -45% |
| Context window | 200K | 200K | — |
| Max output | 8.192K | 128K | — |
Benchmark Comparison
| Benchmark | Claude Haiku 4.5 | GLM-4.7 |
|---|---|---|
| MMLU-Pro | 69.4% | 84.3% |
| HumanEval | 88.1% | — |
| GPQA | — | 85.7% |
Capabilities
| Capability | Claude Haiku 4.5 | GLM-4.7 |
|---|---|---|
| code | ✓ | ✓ |
| reasoning | ✗ | ✓ |
| text | ✓ | ✓ |
| tool-use | ✓ | ✗ |
| vision | ✓ | ✓ |
Claude Haiku 4.5 Strengths
- ✓Very fast responses
- ✓Cheapest Anthropic option
- ✓Good for classification and extraction
Claude Haiku 4.5 Weaknesses
- ✗Weakest reasoning in the Claude family
- ✗Can struggle with nuanced instructions
GLM-4.7 Strengths
- ✓Excellent value — strong benchmarks at $0.60/$2.20
- ✓Open-weight (MIT license)
- ✓Top scores on AIME 25 and BrowseComp
GLM-4.7 Weaknesses
- ✗No tool-use support yet
- ✗358B parameters — still heavy for self-hosting
- ✗Smaller ecosystem than OpenAI/Anthropic
Quick Verdict
Best value: GLM-4.7 is the more affordable option at $0.6/$2.2 per 1M tokens.
Higher benchmarks: GLM-4.7 scores higher on average across available benchmarks (85.0% avg).
Choose GLM-4.7 if cost matters most. Choose Claude Haiku 4.5 if you need the best possible quality for complex tasks.