Run Your Own LLM Evaluation

Test any models with your prompts. Start free, no credit card required.

Start Free

LLM Evaluation

comparative Evaluation

Coding Performance with 10 Evaluators — Run

Comprehensive evaluation of 3 language models across 1 system prompt with rigorous benchmarking and scoring criteria.

Top Score

8.21

glm-5

Average Score

5.00

Spread: 7.31 pts

Avg Latency

—

Response time

Evaluations

120

12 total responses

Executive Insights

Key takeaways from this evaluation

Top Performer

glm-5

8.21

2.31 pts ahead of #2

Model Rankings

Ranked by overall performance score

glm-5

Winner

z-ai/glm-5

Performance Score

8.21/ 10

Very Good

Responses

Avg Latency

—

Cost

$0.0096

qwen3.5-397b-a17b

qwen/qwen3.5-397b-a17b

Performance Score

5.90/ 10

Average

Responses

Avg Latency

—

Cost

$0.0255

llama-4-maverick

meta-llama/llama-4-maverick

Performance Score

0.90/ 10

Needs Improvement

Responses

Avg Latency

—

Cost

$0.0004

Evaluator Consensus

How 10 evaluator models ranked the candidates via blind comparison

majority Agreement

6 of 10 evaluators agree on the top model

glm-5

Avg Rank

1.4

Range

#1–2

#1 Votes

6/10

Latency

—

qwen3.5-397b-a17b

Avg Rank

1.6

Range

#1–2

#1 Votes

4/10

Latency

—

llama-4-maverick

Avg Rank

3.0

Range

#1 Votes

0/10

Latency

—

Per-Evaluator Rankings

How each evaluator model individually ranked the candidates

gpt-5.4-mini

12 evals

glm-58.75

qwen3.5-397b-a17b5.00

llama-4-maverick1.25

gemini-3.1-flash-lite-preview

12 evals

glm-510.00

qwen3.5-397b-a17b3.75

llama-4-maverick1.25

claude-sonnet-4.6

12 evals

glm-58.75

qwen3.5-397b-a17b5.00

llama-4-maverick1.25

minimax-m2.7

12 evals

glm-510.00

qwen3.5-397b-a17b3.75

llama-4-maverick1.25

deepseek-v3.2

12 evals

glm-58.75

qwen3.5-397b-a17b5.00

llama-4-maverick1.25

grok-4.1-fast

12 evals

qwen3.5-397b-a17b7.50

glm-56.25

llama-4-maverick1.25

mistral-small-2603

12 evals

glm-58.75

qwen3.5-397b-a17b6.25

llama-4-maverick0.00

qwen3.5-27b

12 evals

qwen3.5-397b-a17b7.50

glm-57.50

llama-4-maverick0.00

nova-2-lite-v1

12 evals

qwen3.5-397b-a17b7.50

glm-57.50

llama-4-maverick0.00

kimi-k2.5

9 evals

qwen3.5-397b-a17b8.33

glm-55.00

llama-4-maverick1.67

Score Comparison

Visual comparison of all model scores

Run Your Own Model Comparison

Compare any LLM across custom criteria and prompts
Automated scoring with AI evaluators
Share results and track model performance over time

Performance by System Prompt

How each model performs across different evaluation contexts

Coding Agent

12 responses • avg score 5.00

Top Performer

glm-5

8.21

glm-5

8.21

qwen3.5-397b-a17b

5.90

llama-4-maverick

0.90

Performance by Test Prompt

Model results broken down by individual test prompts

Test Prompt	Avg Score	Best	Worst
Javascript Function 3 responses	5.00	9.5 glm-5	0.0 llama-4-maverick
Write an Interval Merge Function 3 responses	5.00	8.0 qwen3.5-397b-a17b	0.0 llama-4-maverick
Debug Python 3 responses	5.00	10.0 glm-5	2.2 qwen3.5-397b-a17b
Refactor Javascript 3 responses	5.00	7.5 qwen3.5-397b-a17b	1.0 llama-4-maverick

About This Evaluation

Methodology, criteria weights, and evaluation confidence

Evaluation Criteria

Method:

comparative

Accuracy50%

Instruction Following50%

Total Responses

120

Total Evaluations