Union Alpha (stealth model) is free for the next week

- no data training
- built for agentic coding
- supports images

let's see what you can do
728
573
11,286
2,574,494
Replying to @opencode
@opencode Any benchmark results? I guess it’s minimax m3.1 or kimi k3.1
7
1
136
84,669
Replying to @theemreates

Sep 16, 2026 · 2:58 PM UTC

68
56
1,340
207,262
RepliesRelevant
Replying to @opencode
@opencode Thank you for sharing guys, Holly f fable level stealth model like wtf it’s obviously new kimi.
5
0
145
18,127
Replying to @opencode
@opencode @theemreates looks like we’ll need a bigger RTX Pro
0
0
1
1,594
Replying to @opencode
@opencode @theemreates What's the point of publishing a benchmark like this given that it's not actually a model, but just whatever random model you got routed to?
0
0
0
226
Replying to @opencode
@opencode @theemreates GLM-5.5 is this you?
0
0
0
931
Replying to @opencode
0
0
34
2,025
Replying to @opencode
@opencode @theemreates If this is real someone has cooked hard
0
0
33
4,063
Replying to @opencode
@opencode @theemreates there is no fucking way this is real
1
0
25
3,407
Replying to @opencode
@opencode @theemreates why didn't you attach this above.
0
0
12
3,203
Replying to @opencode
@opencode @theemreates this is NOT the experience I had
1
0
5
2,541
Replying to @opencode
@opencode @theemreates benchmaxxed for sure
0
0
4
1,641
Replying to @opencode
2
0
3
2,764
Replying to @opencode
@opencode @theemreates Wowwww...
Fable/Astra but significantly cheaper??? 😳😳😳
0
0
3
2,722
Replying to @opencode
@opencode @theemreates what... is this benchmaxxed? I hope not
0
0
3
2,290
Replying to @opencode
@opencode @theemreates cool but what
0
0
2
1,104
Replying to @opencode
@opencode @theemreates Why is not deepseek v4.1 flash in this picture?
1
0
1
641
Replying to @opencode
@opencode @theemreates So that's Jev?
0
0
1
75
Replying to @opencode
@opencode @theemreates That's impressive. What about the other benchmarks?
0
0
1
2,238
Replying to @opencode
@opencode @theemreates Hello.
what ?

this looks so good
0
0
1
3,695
Replying to @opencode
0
0
1
919
Replying to @opencode
@opencode @theemreates wait, what the heck?
0
0
1
2,424
Replying to @opencode
@opencode @theemreates Oh, It makes new pareto line
0
0
1
3,076
Replying to @opencode
@opencode @theemreates Deepseek v4.1 flash?
0
0
0
198
Replying to @opencode
@opencode @theemreates not seeing this model on linux
0
0
0
905
Replying to @opencode
@opencode @theemreates 评测集比 demo 视频更能证明你交付过。
0
0
0
961
Replying to @opencode
@opencode @theemreates did someone try already?
1
0
0
2,132
Replying to @opencode
@opencode @theemreates damn if it is sloooooow
0
0
0
788
Replying to @opencode
@opencode @theemreates so it's close to Astra Medium or Astra Low then?
0
0
0
3,617
Replying to @opencode
@opencode @theemreates How it is related to omen alpha?
0
0
0
4,971
Replying to @opencode
0
0
1
808
Replying to @opencode
0
0
0
100
Replying to @opencode
@opencode @theemreates cc @tautologer no way right? Like there’s no chance

@_ontologic @somewheresy
1
0
2
791
Replying to @opencode
0
0
1
261
Replying to @opencode
@opencode @theemreates 看起来性价比可以啊
0
0
1
373
Replying to @opencode
@opencode @theemreates Welcome Composer 3.0
0
0
1
427
Replying to @opencode
0
0
0
144
Replying to @opencode
@opencode @theemreates Thank you. It's amazing, and didn't expect you have this chart ahead of the launch or testing.
0
0
1
1,563