# Claude 3 Haiku vs GLM-5

> GLM-5 is the stronger model overall, scoring 46.1 to 25.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-3-haiku-vs-glm-5
- Last updated: 2026-10-10
- Shared benchmarks: 24

## Summary

- They share 24 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and GLM-5 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GLM-5 leads 46.4 to 9.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.8% for Claude 3 Haiku and 80% for GLM-5.
- GLM-5 has downloadable open weights; the other is API-only.

## Snapshot

| | Claude 3 Haiku | GLM-5 |
|---|---|---|
| Provider | Anthropic | Z.ai (Zhipu) |
| Noometry Index | 25.9 | 46.1 |
| Rank | 340 | 66 |
| Context | — | 205K |
| Input $/M | — | $1 |
| Output $/M | — | $3.20 |
| Weights | Proprietary | Open |

## Coding

- Claude 3 Haiku: 26.4 (#325)
- GLM-5: 49.0 (#52)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| WeirdML | 9.8% | 48.2% |
| LMArena Coding | 1199 | 1461 |
| SWE-bench Verified | — | 72.1% |
| SWE-bench Verified (bash only) | — | 72.8% |
| LMArena WebDev | — | 1434 |
| SWE-bench Multilingual | — | 69.7% |
| BigCodeBench Instruct | 39.4% | — |
| BigCodeBench Complete | 50.1% | — |
| CadEval | 12% | — |
| ALE-Bench | — | 765.62 |
| HumanEval+ | 68.9% | — |
| MBPP+ | 68.8% | — |

## Agentic & Tool Use

- Claude 3 Haiku: —
- GLM-5: 31.1 (#71)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| Terminal-Bench | — | 52.4% |
| τ²-bench Airline | — | 82.5% |
| τ²-bench Banking | — | 9.8% |
| τ²-bench Retail | — | 73.7% |
| τ²-bench Telecom | — | 86.8% |
| Vending-Bench 2 | — | 4,432 |

## Reasoning

- Claude 3 Haiku: 16.3 (#307)
- GLM-5: 27.6 (#116)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| Kagi LLM Benchmark | 34.2% | 75% |
| LMArena Hard Prompts | 1174 | 1452 |
| Epoch Capabilities Index | 118.35 | 145.83 |
| ForecastBench | 53.2 | 61 |
| ARC-AGI-2 | — | 4.9% |
| SimpleBench | — | 53.2% |
| NYT Connections (extended) | — | 74.8% |
| ARC-AGI-1 | — | 44.7% |
| Chess Puzzles | — | 10% |
| DTBench | 50.1% | — |
| LMCA | 8.8% | — |
| WinoGrande | 74.2% | — |

## Math

- Claude 3 Haiku: 9.8 (#319)
- GLM-5: 46.4 (#71)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 80% |
| LMArena Math | 1188 | 1440 |
| MathArena Final-Answer Competitions | — | 65.7% |
| MATH Level 5 | 14.9% | — |
| FrontierMath (Feb 2025 set) | — | 16.4% |
| FrontierMath Tier 4 (v1) | — | 2.1% |

## Knowledge

- Claude 3 Haiku: 17.3 (#285)
- GLM-5: 52.3 (#64)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| GPQA Diamond | 36.3% | 87.8% |
| LMArena Expert | 1148 | 1454 |
| Confabulations | 34.2% | — |
| Vectara Hallucination Rate | — | 10.1% |
| MMLU | 73.8% | — |

## Multimodal

- Claude 3 Haiku: 23.6 (#128)
- GLM-5: —

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| LMArena Vision | 950 | — |
| ScienceQA | 72% | — |

## Multilingual

- Claude 3 Haiku: 36.0 (#243)
- GLM-5: 53.7 (#58)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| LMArena Non-English | 1178 | 1430 |
| LMArena Chinese | 1155 | 1511 |
| LMArena French | 1195 | 1455 |
| LMArena German | 1174 | 1445 |
| LMArena Japanese | 1102 | 1416 |
| LMArena Korean | 1109 | 1423 |
| LMArena Russian | 1204 | 1436 |
| LMArena Spanish | 1166 | 1454 |

## Instruction Following

- Claude 3 Haiku: 61.3 (#247)
- GLM-5: 75.2 (#67)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| LMArena Instruction Following | 1173 | 1428 |

## Long Context

- Claude 3 Haiku: 36.1 (#237)
- GLM-5: 44.7 (#60)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| LMArena Longer Query | 1190 | 1446 |
| CL-bench | — | 18.7% |

## Writing & Preference

- Claude 3 Haiku: 29.7 (#291)
- GLM-5: 66.0 (#38)

| Benchmark | Claude 3 Haiku | GLM-5 |
|---|---|---|
| LMArena Text | 1195 | 1446 |
| LMArena Creative Writing | 1157 | 1439 |
| EQ-Bench Creative Writing | 717 | 1601 |
| LMArena Multi-Turn | 1190 | 1456 |

## FAQ

### Is Claude 3 Haiku better than GLM-5?

GLM-5 is the stronger model overall, scoring 46.1 to 25.9 on the Noometry Index.

### Is Claude 3 Haiku or GLM-5 better for coding?

GLM-5 scores higher on coding benchmarks: 49.0 versus 26.4 in the Noometry coding category.

### How many benchmarks do Claude 3 Haiku and GLM-5 share?

24 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and GLM-5 has 45.