TLDR: Probably the best gguf version, beating unsloth UD.

Best use case
Function Calling = reasoning OFF
Multimodal = reasoning ON

📊 Configuration:
Base URL: http://localhost:9099/v1
Model: Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q5_K_P_off
Test Cases: 17
Output: results/agent_test_results_Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q5_K_P_off_20260405_153735.json
Log File: logs/agent_test_logs_Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q5_K_P_off_20260405_153735.log

🔄 Running agent tests...
Starting agent test suite with 17 test cases
Running agent test: zero_greeting
Running agent test: zero_weather_question
Running agent test: zero_general_question
Running agent test: simple_checkout
Running agent test: zero_thank_you
Running agent test: medium_search_and_add
Running agent test: zero_capabilities
Running agent test: simple_search_electronics
Running agent test: simple_add_iphone
Running agent test: simple_view_cart
Running agent test: complex_shopping_workflow
Running agent test: medium_remove_and_add
Running agent test: medium_view_and_add
Running agent test: complex_cart_management
Running agent test: complex_gift_shopping
Running agent test: medium_search_category_and_add
Running agent test: simple_remove_product
✅ Tests completed in 29.6192272s

📈 Agent Test Results

Total Tests: 17
✅ Passed: 16
❌ Failed: 1
⏱️ Total LLM Time: 4m56.6324561s
⏱️ Average Time per Request: 7.234937953s

📋 Test Case Results:

Test Case: zero_general_question
Status: ✅ PASSED
Matched Path: no_tools
Response Time: 7.1037635s
Tool Calls: 0

Test Case: zero_greeting
Status: ✅ PASSED
Matched Path: no_tools
Response Time: 7.6242963s
Tool Calls: 0

Test Case: zero_thank_you
Status: ✅ PASSED
Matched Path: no_tools
Response Time: 8.6742927s
Tool Calls: 0

Test Case: zero_weather_question
Status: ✅ PASSED
Matched Path: no_tools
Response Time: 10.222154s
Tool Calls: 0

Test Case: zero_capabilities
Status: ✅ PASSED
Matched Path: no_tools
Response Time: 10.7087085s
Tool Calls: 0

Test Case: simple_view_cart
Status: ✅ PASSED
Matched Path: view_cart
Response Time: 11.081606s
Tool Calls: 1
Tools Used: view_cart

Test Case: simple_remove_product
Status: ✅ PASSED
Matched Path: direct_remove
Response Time: 14.1713419s
Tool Calls: 1
Tools Used: remove_from_cart

Test Case: simple_checkout
Status: ✅ PASSED
Matched Path: direct_checkout
Response Time: 16.2663444s
Tool Calls: 1
Tools Used: checkout

Test Case: simple_add_iphone
Status: ✅ PASSED
Matched Path: direct_add
Response Time: 16.8710434s
Tool Calls: 1
Tools Used: add_to_cart

Test Case: simple_search_electronics
Status: ✅ PASSED
Matched Path: search_by_category
Response Time: 18.9555705s
Tool Calls: 1
Tools Used: search_products

Test Case: medium_search_and_add
Status: ✅ PASSED
Matched Path: search_by_query
Response Time: 20.6402634s
Tool Calls: 2
Tools Used: search_products, add_to_cart

Test Case: medium_remove_and_add
Status: ✅ PASSED
Matched Path: remove_then_add
Response Time: 21.4799744s
Tool Calls: 2
Tools Used: remove_from_cart, add_to_cart

Test Case: medium_search_category_and_add
Status: ✅ PASSED
Matched Path: search_then_add
Response Time: 23.1171036s
Tool Calls: 2
Tools Used: search_products, add_to_cart

Test Case: complex_gift_shopping
Status: ✅ PASSED
Matched Path: gift_shopping_workflow
Response Time: 24.8069479s
Tool Calls: 5
Tools Used: search_products, add_to_cart, search_products, add_to_cart, view_cart

Test Case: medium_view_and_add
Status: ✅ PASSED
Matched Path: view_then_add
Response Time: 26.8041455s
Tool Calls: 2
Tools Used: view_cart, add_to_cart

Test Case: complex_shopping_workflow
Status: ❌ FAILED
Response Time: 28.7320275s
Tool Calls: 5
Tools Used: search_products, add_to_cart, add_to_cart, add_to_cart, checkout

Test Case: complex_cart_management
Status: ✅ PASSED
Matched Path: cart_organization
Response Time: 29.617127s
Tool Calls: 3
Tools Used: view_cart, remove_from_cart, add_to_cart

❌ Failed Tests Details:

Test Case: complex_shopping_workflow
Expected Tool Variants: 4
Variant 1 (full_workflow_with_iphone): 4 tools
Variant 2 (full_workflow_with_headphones): 4 tools
Variant 3 (full_workflow_with_headphones_and_iphone): 5 tools
Variant 4 (full_workflow_with_iphone_and_headphones): 5 tools
Actual Tool Calls: 5
1. search_products
2. add_to_cart
3. add_to_cart
4. add_to_cart
5. checkout
Response Time: 28.7320275s

📊 Overall Success Rate: 94.12% // https://github.com/docker/model-test/

Q5_K_P // 60 tokens/s on RTX 3060 12GB (reasoning OFF)

⏱️ Average Time per Request: 7.234937953s
📊 Overall Success Rate: 94.12%

Q5_K_P // 60 tokens/s on RTX 3060 12GB (reasoning ON)

⏱️ Average Time per Request: 19.598046543s
📊 Overall Success Rate: 88.24%

Function Calling = reasoning OFF
Multimodal = reasoning ON

llama-server --host 0.0.0.0 --port 9099 -ngl 99 -fa on -c 65536 --temp 1 --top-k 64 --top-p 0.95 --jinja -m X:\path\to\Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q5_K_P.gguf --mmproj X:\path\to\mmproj-Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-f16.gguf --reasoning on

TLDR: Probably the best gguf version, beating unsloth UD.