# GLM-4.6 vs Llama 3-8B

> GLM-4.6 is the stronger model overall, scoring 40.9 to 25.3 on the Noometry Index. Llama 3-8B costs 7.1× less per token, which makes it the better buy when GLM-4.6's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/glm-4-6-vs-llama-3-8b
- Last updated: 2026-10-10
- Shared benchmarks: 17

## Summary

- They share 17 benchmarks with published results for both. GLM-4.6 scores higher in 8 categories and Llama 3-8B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GLM-4.6 leads 39.6 to 7.8.
- Llama 3-8B is cheaper at $0.14 / $0.14 per million input/output tokens, against $0.60 / $2.20 for GLM-4.6.
- GLM-4.6 accepts more context: 205K tokens versus 8K.

## Snapshot

| | GLM-4.6 | Llama 3-8B |
|---|---|---|
| Provider | Z.ai (Zhipu) | Meta |
| Noometry Index | 40.9 | 25.3 |
| Rank | 132 | 338 |
| Context | 205K | 8K |
| Input $/M | $0.60 | $0.14 |
| Output $/M | $2.20 | $0.14 |
| Weights | Open | Open |

## Coding

- GLM-4.6: 39.9 (#137)
- Llama 3-8B: 30.5 (#285)

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| LMArena Coding | 1449 | 1152 |
| SWE-bench Verified (bash only) | 55.4% | — |
| LMArena WebDev | 1339 | — |
| SciCode | 38.4% | — |
| BigCodeBench Instruct | — | 31.9% |
| BigCodeBench Complete | — | 36.9% |
| ALE-Bench | 340.82 | — |
| HumanEval+ | — | 56.7% |
| MBPP+ | — | 54.8% |

## Agentic & Tool Use

- GLM-4.6: 32.1 (#60)
- Llama 3-8B: —

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| Terminal-Bench | 24.5% | — |
| Berkeley Function Calling Leaderboard | 72.4% | — |

## Reasoning

- GLM-4.6: 23.1 (#168)
- Llama 3-8B: 13.8 (#321)

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| LMArena Hard Prompts | 1440 | 1133 |
| Kagi LLM Benchmark | 47.4% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | — | 0% |
| DTBench | — | 43.9% |
| Adversarial NLI | — | 57.3% |
| Epoch Capabilities Index | — | 116.33 |
| ForecastBench | — | 58.6 |
| WinoGrande | — | 75.7% |

## Math

- GLM-4.6: 36.4 (#115)
- Llama 3-8B: 8.3 (#317)

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| LMArena Math | 1431 | 1151 |
| OTIS Mock AIME 2024-2025 | — | 1.9% |
| MATH Level 5 | — | 6.1% |
| FrontierMath (Feb 2025 set) | 3.8% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |

## Knowledge

- GLM-4.6: 39.6 (#118)
- Llama 3-8B: 7.8 (#303)

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| LMArena Expert | 1432 | 1113 |
| GPQA Diamond | — | 26.1% |
| Vectara Hallucination Rate | 9.5% | — |
| ARC (AI2) Challenge | — | 82.8% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 82.6% |
| TriviaQA | — | 67.7% |

## Multilingual

- GLM-4.6: 53.6 (#62)
- Llama 3-8B: 31.0 (#256)

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| LMArena Non-English | 1426 | 1098 |
| LMArena Chinese | 1500 | 1076 |
| LMArena French | 1459 | 1158 |
| LMArena German | 1447 | 1104 |
| LMArena Japanese | 1394 | 967 |
| LMArena Korean | 1401 | 1003 |
| LMArena Russian | 1418 | 1109 |
| LMArena Spanish | 1436 | 1172 |

## Instruction Following

- GLM-4.6: 74.5 (#93)
- Llama 3-8B: 58.7 (#255)

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| LMArena Instruction Following | 1410 | 1127 |

## Long Context

- GLM-4.6: 43.4 (#91)
- Llama 3-8B: 34.2 (#246)

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| LMArena Longer Query | 1423 | 1128 |

## Writing & Preference

- GLM-4.6: 61.3 (#86)
- Llama 3-8B: 37.7 (#251)

| Benchmark | GLM-4.6 | Llama 3-8B |
|---|---|---|
| LMArena Text | 1440 | 1166 |
| LMArena Creative Writing | 1411 | 1150 |
| LMArena Multi-Turn | 1427 | 1152 |
| EQ-Bench Creative Writing | 1411 | — |

## FAQ

### Is GLM-4.6 better than Llama 3-8B?

GLM-4.6 is the stronger model overall, scoring 40.9 to 25.3 on the Noometry Index. Llama 3-8B costs 7.1× less per token, which makes it the better buy when GLM-4.6's lead doesn't matter for your workload.

### Which is cheaper, GLM-4.6 or Llama 3-8B?

Llama 3-8B is cheaper. It lists at $0.14 per million input tokens and $0.14 per million output tokens; GLM-4.6 lists at $0.60 and $2.20.

### Is GLM-4.6 or Llama 3-8B better for coding?

GLM-4.6 scores higher on coding benchmarks: 39.9 versus 30.5 in the Noometry coding category.

### Which has the bigger context window?

GLM-4.6 does, with 205K tokens against 8K.

### How many benchmarks do GLM-4.6 and Llama 3-8B share?

17 benchmarks have published results for both models. GLM-4.6 has 29 scored results on Noometry and Llama 3-8B has 34.