|
39 | 39 | "upvote_ratio": 0.83, |
40 | 40 | "author_flair_background_color": null, |
41 | 41 | "subreddit_type": "public", |
42 | | - "ups": 59, |
| 42 | + "ups": 60, |
43 | 43 | "total_awards_received": 0, |
44 | 44 | "media_embed": {}, |
45 | 45 | "thumbnail_width": null, |
|
53 | 53 | "secure_media_embed": {}, |
54 | 54 | "link_flair_text": "Discussion", |
55 | 55 | "can_mod_post": false, |
56 | | - "score": 59, |
| 56 | + "score": 60, |
57 | 57 | "approved_by": null, |
58 | 58 | "is_created_from_ads_ui": false, |
59 | 59 | "author_premium": false, |
|
116 | 116 | "permalink": "/r/LocalLLaMA/comments/1mg3d62/note_to_the_qwen_team_re_the_new_30b_a3b_coder/", |
117 | 117 | "stickied": false, |
118 | 118 | "url": "https://www.reddit.com/r/LocalLLaMA/comments/1mg3d62/note_to_the_qwen_team_re_the_new_30b_a3b_coder/", |
119 | | - "subreddit_subscribers": 509626, |
| 119 | + "subreddit_subscribers": 509919, |
120 | 120 | "created_utc": 1754174302, |
121 | 121 | "num_crossposts": 0, |
122 | 122 | "mod_reports": [], |
|
165 | 165 | "created_utc": 1754180230, |
166 | 166 | "send_replies": true, |
167 | 167 | "parent_id": "t3_1mg3d62", |
168 | | - "score": 36, |
| 168 | + "score": 37, |
169 | 169 | "author_fullname": "t2_98d256k", |
170 | 170 | "approved_by": null, |
171 | 171 | "mod_note": null, |
|
207 | 207 | "collapsed_because_crowd_control": null, |
208 | 208 | "mod_reports": [], |
209 | 209 | "num_reports": null, |
210 | | - "ups": 36 |
| 210 | + "ups": 37 |
211 | 211 | } |
212 | 212 | }, |
213 | 213 | { |
|
831 | 831 | "created_utc": 1754184900, |
832 | 832 | "send_replies": true, |
833 | 833 | "parent_id": "t3_1mg3d62", |
834 | | - "score": 15, |
| 834 | + "score": 16, |
835 | 835 | "author_fullname": "t2_w4j8t", |
836 | 836 | "approved_by": null, |
837 | 837 | "mod_note": null, |
|
873 | 873 | "collapsed_because_crowd_control": null, |
874 | 874 | "mod_reports": [], |
875 | 875 | "num_reports": null, |
876 | | - "ups": 15 |
| 876 | + "ups": 16 |
877 | 877 | } |
878 | 878 | }, |
879 | 879 | { |
|
912 | 912 | "mod_note": null, |
913 | 913 | "all_awardings": [], |
914 | 914 | "collapsed": false, |
915 | | - "body": "It must vary depending on the task. For me, Qwen 3 Coder 30B-A3B has been excellent, and substantially better than the non-coder 2507 Instruct model. While the coder model’s general world knowledge is a lot worse as expected, its knowledge of language features, APIs, popular APIs, and behaviour of APIs is vastly superior to the non-coder for me. Coder has also done substantially better for me at generating working code and understanding large complex codebases.\n\nI’ve been using it for embedded C and Rust things, and some Python web service backend stuff.", |
916 | | - "edited": false, |
| 915 | + "body": "It must vary depending on the task. For me, Qwen 3 Coder 30B-A3B has been excellent, and substantially better than the non-coder 2507 Instruct model. While the coder model’s general world knowledge is a lot worse as expected, its knowledge of language features, popular libraries/APIs, and behaviour of APIs is vastly superior to the non-coder for me. Coder has also done substantially better for me at generating working code and understanding large complex codebases.\n\nI’ve been using it for embedded C and Rust things, and some Python web service backend stuff.", |
| 916 | + "edited": 1754287574, |
917 | 917 | "top_awarded_type": null, |
918 | 918 | "author_flair_css_class": null, |
919 | 919 | "name": "t1_n6mwu1o", |
920 | 920 | "is_submitter": false, |
921 | 921 | "downs": 0, |
922 | 922 | "author_flair_richtext": [], |
923 | 923 | "author_patreon_flair": false, |
924 | | - "body_html": "<div class=\"md\"><p>It must vary depending on the task. For me, Qwen 3 Coder 30B-A3B has been excellent, and substantially better than the non-coder 2507 Instruct model. While the coder model’s general world knowledge is a lot worse as expected, its knowledge of language features, APIs, popular APIs, and behaviour of APIs is vastly superior to the non-coder for me. Coder has also done substantially better for me at generating working code and understanding large complex codebases.</p>\n\n<p>I’ve been using it for embedded C and Rust things, and some Python web service backend stuff.</p>\n</div>", |
| 924 | + "body_html": "<div class=\"md\"><p>It must vary depending on the task. For me, Qwen 3 Coder 30B-A3B has been excellent, and substantially better than the non-coder 2507 Instruct model. While the coder model’s general world knowledge is a lot worse as expected, its knowledge of language features, popular libraries/APIs, and behaviour of APIs is vastly superior to the non-coder for me. Coder has also done substantially better for me at generating working code and understanding large complex codebases.</p>\n\n<p>I’ve been using it for embedded C and Rust things, and some Python web service backend stuff.</p>\n</div>", |
925 | 925 | "removal_reason": null, |
926 | 926 | "collapsed_reason": null, |
927 | 927 | "distinguished": null, |
|
981 | 981 | "created_utc": 1754211001, |
982 | 982 | "send_replies": true, |
983 | 983 | "parent_id": "t3_1mg3d62", |
984 | | - "score": 6, |
| 984 | + "score": 4, |
985 | 985 | "author_fullname": "t2_q1q7nlxa", |
986 | 986 | "approved_by": null, |
987 | 987 | "mod_note": null, |
|
1023 | 1023 | "collapsed_because_crowd_control": null, |
1024 | 1024 | "mod_reports": [], |
1025 | 1025 | "num_reports": null, |
1026 | | - "ups": 6 |
| 1026 | + "ups": 4 |
1027 | 1027 | } |
1028 | 1028 | }, |
1029 | 1029 | { |
|
1131 | 1131 | "created_utc": 1754194177, |
1132 | 1132 | "send_replies": true, |
1133 | 1133 | "parent_id": "t3_1mg3d62", |
1134 | | - "score": 2, |
| 1134 | + "score": 1, |
1135 | 1135 | "author_fullname": "t2_yrl9ztfsa", |
1136 | 1136 | "approved_by": null, |
1137 | 1137 | "mod_note": null, |
|
1173 | 1173 | "collapsed_because_crowd_control": null, |
1174 | 1174 | "mod_reports": [], |
1175 | 1175 | "num_reports": null, |
1176 | | - "ups": 2 |
| 1176 | + "ups": 1 |
1177 | 1177 | } |
1178 | 1178 | }, |
1179 | 1179 | { |
|
1642 | 1642 | "can_mod_post": false, |
1643 | 1643 | "send_replies": true, |
1644 | 1644 | "parent_id": "t1_n6lw3do", |
1645 | | - "score": 9, |
| 1645 | + "score": 8, |
1646 | 1646 | "author_fullname": "t2_1mhx5", |
1647 | 1647 | "removal_reason": null, |
1648 | 1648 | "approved_by": null, |
|
1685 | 1685 | "collapsed_because_crowd_control": null, |
1686 | 1686 | "mod_reports": [], |
1687 | 1687 | "num_reports": null, |
1688 | | - "ups": 9 |
| 1688 | + "ups": 8 |
1689 | 1689 | } |
1690 | 1690 | } |
1691 | 1691 | ], |
|
1706 | 1706 | "created_utc": 1754177363, |
1707 | 1707 | "send_replies": true, |
1708 | 1708 | "parent_id": "t1_n6lqwd2", |
1709 | | - "score": 6, |
| 1709 | + "score": 7, |
1710 | 1710 | "author_fullname": "t2_8eneodlk", |
1711 | 1711 | "removal_reason": null, |
1712 | 1712 | "approved_by": null, |
|
1753 | 1753 | "collapsed_because_crowd_control": null, |
1754 | 1754 | "mod_reports": [], |
1755 | 1755 | "num_reports": null, |
1756 | | - "ups": 6 |
| 1756 | + "ups": 7 |
1757 | 1757 | } |
1758 | 1758 | }, |
1759 | 1759 | { |
|
2332 | 2332 | "created_utc": 1754196779, |
2333 | 2333 | "send_replies": true, |
2334 | 2334 | "parent_id": "t3_1mg3d62", |
2335 | | - "score": 3, |
| 2335 | + "score": 1, |
2336 | 2336 | "author_fullname": "t2_1qychuraq9", |
2337 | 2337 | "approved_by": null, |
2338 | 2338 | "mod_note": null, |
|
2374 | 2374 | "collapsed_because_crowd_control": null, |
2375 | 2375 | "mod_reports": [], |
2376 | 2376 | "num_reports": null, |
2377 | | - "ups": 3 |
| 2377 | + "ups": 1 |
2378 | 2378 | } |
2379 | 2379 | } |
2380 | 2380 | ], |
|
0 commit comments