Skip to content

Instantly share code, notes, and snippets.

@anzax
Created April 16, 2026 09:21
Show Gist options
  • Select an option

  • Save anzax/1d6501e115d8d1208121fb697736fe12 to your computer and use it in GitHub Desktop.

Select an option

Save anzax/1d6501e115d8d1208121fb697736fe12 to your computer and use it in GitHub Desktop.
aider bench runs Qwen3.5-122B-A10B
3. eugr docker - v2 autoround
- dirname: 2026-04-13-13-18-52--qwen35-122b-shieldstar-autoround-vllm-full-n4
test_cases: 225
model: openai/Intel/Qwen3.5-122B-A10B-int4-AutoRound
edit_format: diff
commit_hash: f09d706
pass_rate_1: 37.8
pass_rate_2: 71.6
pass_num_1: 85
pass_num_2: 161
percent_cases_well_formed: 94.2
error_outputs: 16
num_malformed_responses: 15
num_with_malformed_responses: 13
user_asks: 76
lazy_comments: 20
syntax_errors: 0
indentation_errors: 0
exhausted_context_windows: 1
prompt_tokens: 3207814
completion_tokens: 591014
test_timeouts: 0
total_tests: 225
command: aider --model openai/Intel/Qwen3.5-122B-A10B-int4-AutoRound
date: 2026-04-13
versions: 0.86.3.dev
seconds_per_case: 200.6
—-
4. QuantTrio/Qwen3.5-122B-A10B-AWQ
- dirname: 2026-04-13-19-36-44--qwen35-122b-AWQ-vllm-full-n4
test_cases: 225
model: openai/QuantTrio/Qwen3.5-122B-A10B-AWQ
edit_format: diff
commit_hash: f09d706
pass_rate_1: 42.7
pass_rate_2: 75.1
pass_num_1: 96
pass_num_2: 169
percent_cases_well_formed: 97.8
error_outputs: 7
num_malformed_responses: 5
num_with_malformed_responses: 5
user_asks: 76
lazy_comments: 1
syntax_errors: 0
indentation_errors: 0
exhausted_context_windows: 1
prompt_tokens: 2703388
completion_tokens: 515893
test_timeouts: 0
total_tests: 225
command: aider --model openai/QuantTrio/Qwen3.5-122B-A10B-AWQ
date: 2026-04-13
versions: 0.86.3.dev
seconds_per_case: 206.6
5.
- dirname: 2026-04-14-00-15-20--qwen35-122b-AWQ-cyankiwi-vllm-full-n4
test_cases: 225
model: openai/cyankiwi/Qwen3.5-122B-A10B-AWQ-4bit
edit_format: diff
commit_hash: f09d706
pass_rate_1: 44.0
pass_rate_2: 72.0
pass_num_1: 99
pass_num_2: 162
percent_cases_well_formed: 93.3
error_outputs: 18
num_malformed_responses: 17
num_with_malformed_responses: 15
user_asks: 78
lazy_comments: 2
syntax_errors: 0
indentation_errors: 0
exhausted_context_windows: 1
prompt_tokens: 2988059
completion_tokens: 558452
test_timeouts: 0
total_tests: 225
command: aider --model openai/cyankiwi/Qwen3.5-122B-A10B-AWQ-4bit
date: 2026-04-14
versions: 0.86.3.dev
seconds_per_case: 254.4
6.
- dirname: 2026-04-14-09-15-56--qwen35-122b-AWQ-cyankiwi-vllm-full-n4
test_cases: 225
model: openai/cyankiwi/Qwen3.5-122B-A10B-AWQ-4bit
edit_format: diff
commit_hash: f09d706
pass_rate_1: 39.6
pass_rate_2: 74.2
pass_num_1: 89
pass_num_2: 167
percent_cases_well_formed: 92.0
error_outputs: 20
num_malformed_responses: 18
num_with_malformed_responses: 18
user_asks: 77
lazy_comments: 2
syntax_errors: 0
indentation_errors: 0
exhausted_context_windows: 1
prompt_tokens: 3194640
completion_tokens: 575788
test_timeouts: 1
total_tests: 225
command: aider --model openai/cyankiwi/Qwen3.5-122B-A10B-AWQ-4bit
date: 2026-04-14
versions: 0.86.3.dev
seconds_per_case: 266.3
——
7.
- dirname: 2026-04-14-14-06-57--qwen35-122b-AWQ-QuantTrio-vllm-full-n4
test_cases: 225
model: openai/QuantTrio/Qwen3.5-122B-A10B-AWQ
edit_format: diff
commit_hash: f09d706
pass_rate_1: 42.7
pass_rate_2: 74.7
pass_num_1: 96
pass_num_2: 168
percent_cases_well_formed: 91.1
error_outputs: 23
num_malformed_responses: 22
num_with_malformed_responses: 20
user_asks: 69
lazy_comments: 3
syntax_errors: 0
indentation_errors: 0
exhausted_context_windows: 1
prompt_tokens: 2602444
completion_tokens: 544141
test_timeouts: 1
total_tests: 225
command: aider --model openai/QuantTrio/Qwen3.5-122B-A10B-AWQ
date: 2026-04-14
versions: 0.86.3.dev
seconds_per_case: 215.4
8.
- dirname: 2026-04-14-20-17-56--qwen35-122b-int4-hybrid-vllm-full-n6
test_cases: 225
model: openai/Qwen3.5-122B-A10B-int4-AutoRound
edit_format: diff
commit_hash: f09d706
pass_rate_1: 37.3
pass_rate_2: 70.2
pass_num_1: 84
pass_num_2: 158
percent_cases_well_formed: 94.2
error_outputs: 17
num_malformed_responses: 16
num_with_malformed_responses: 13
user_asks: 76
lazy_comments: 2
syntax_errors: 0
indentation_errors: 0
exhausted_context_windows: 1
prompt_tokens: 3115980
completion_tokens: 564114
test_timeouts: 1
total_tests: 225
command: aider --model openai/Qwen3.5-122B-A10B-int4-AutoRound
date: 2026-04-14
versions: 0.86.3.dev
seconds_per_case: 239.9
total_cost: 0.0000
9.
- dirname: 2026-04-14-23-59-13--qwen35-122b-awq-fp8kv-vllm-full-n6
test_cases: 225
model: openai/QuantTrio/Qwen3.5-122B-A10B-AWQ
edit_format: diff
commit_hash: f09d706
pass_rate_1: 44.4
pass_rate_2: 74.7
pass_num_1: 100
pass_num_2: 168
percent_cases_well_formed: 94.2
error_outputs: 18
num_malformed_responses: 15
num_with_malformed_responses: 13
user_asks: 80
lazy_comments: 2
syntax_errors: 0
indentation_errors: 0
exhausted_context_windows: 1
prompt_tokens: 2477802
completion_tokens: 505004
test_timeouts: 0
total_tests: 225
command: aider --model openai/QuantTrio/Qwen3.5-122B-A10B-AWQ
date: 2026-04-14
versions: 0.86.3.dev
seconds_per_case: 243.5
total_cost: 0.0000
10.
- dirname: 2026-04-15-07-46-50--qwen35-122b-awq-fp8kv-vllm-full-n6
test_cases: 225
model: openai/QuantTrio/Qwen3.5-122B-A10B-AWQ
edit_format: diff
commit_hash: f09d706
pass_rate_1: 40.0
pass_rate_2: 75.6
pass_num_1: 90
pass_num_2: 170
percent_cases_well_formed: 92.4
error_outputs: 18
num_malformed_responses: 17
num_with_malformed_responses: 17
user_asks: 92
lazy_comments: 3
syntax_errors: 0
indentation_errors: 0
exhausted_context_windows: 1
prompt_tokens: 2886931
completion_tokens: 557105
test_timeouts: 1
total_tests: 225
command: aider --model openai/QuantTrio/Qwen3.5-122B-A10B-AWQ
date: 2026-04-15
versions: 0.86.3.dev
seconds_per_case: 248.8
total_cost: 0.0000
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment