Request a pilot
‹ PublicAI Index

The LLM benchmark aggregator.

Union Alpha

Other

Strongest in Language (#9 of 57), weakest in Core abilities (#88 of 204). Above par in 9 of 11 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and Claude Fable 5.1 and ahead of Qwen3.8 Flash Next and Claude Opus 4.6.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Coding54.6#34/165
Code generation59.6#10/77
Agentic coding51.4#62/157
Reasoning52.4#71/178
Mathematics59#13/140
Reasoning45.8#77/140
Core abilities50.5#88/204
Language62.5#9/57
Data analysis52#30/57
Instruction following35.2#51/57
General intelligence52.1#79/204

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Union Alpha, left for the other.

Granite 4.2 3Bwon 7 · lost 0
GPT-4.1 Nanowon 7 · lost 0
Granite 4.2 8Bwon 7 · lost 0
Mistral Large 3won 7 · lost 0
Gemini 2.0 Flashwon 7 · lost 0
K2 Horizon 3.7Bwon 7 · lost 0
Claude Haiku 4.5won 7 · lost 0
GPT OSS 120Bwon 7 · lost 0
GPT-4.1 Miniwon 7 · lost 0
DeepSeek V3.2won 7 · lost 0
GPT-4o Miniwon 8 · lost 0
Qwen3.6 27Bwon 11 · lost 0
Qwen3.6 Pluswon 11 · lost 0
Grok 4.3won 10 · lost 1
Gemini 3.5 Flash Litewon 10 · lost 1
GPT-5.4 Miniwon 10 · lost 1
GPT-5.4 Nanowon 10 · lost 1
Kimi K2.7 Codewon 10 · lost 1
Kimi K2.6won 10 · lost 1
Grok Build 0.1won 10 · lost 1
MiniMax M3won 10 · lost 1
DeepSeek V3won 7 · lost 1
K2 Horizon 7Bwon 6 · lost 1
Nemotron 3 Ultrawon 6 · lost 1
Kimi K2won 6 · lost 1
GPT-4.1won 6 · lost 1
MiniMax M2.5won 6 · lost 1
Ox Alphawon 9 · lost 2
Inklingwon 9 · lost 2
Gemini 2.5 Flashwon 6 · lost 2
GPT-6 Lunawon 8 · lost 3
GLM-5.3 Flashwon 8 · lost 3
GPT-5.2 Codexwon 8 · lost 3
Gemini 3.6 Flashwon 8 · lost 3
Grok 3 Miniwon 5 · lost 2
Kimi K2 Thinkingwon 5 · lost 2
GPT-5 Miniwon 5 · lost 2
GLM-5.2won 7 · lost 4
Claude Sonnet 4.6won 7 · lost 4
Qwen3.7 Maxwon 7 · lost 4
Gemini 3.5 Flashwon 7 · lost 4
Claude Sonnet 4won 5 · lost 3
Claude Sonnet 4.5won 4 · lost 3
Grok 4 Fastwon 4 · lost 3
GLM-5won 4 · lost 3
Kimi K2.5won 4 · lost 3
Muse Spark 1.1won 6 · lost 5
Claude Opus 4.5won 6 · lost 5
Claude Opus 4.6won 6 · lost 5
Qwen3.8 Flash Nextwon 6 · lost 5
DeepSeek R1won 4 · lost 4
Gemini 2.5 Prowon 4 · lost 4
Claude Opus 4won 4 · lost 4
Claude Sonnet 5won 5 · lost 6
DeepSeek V4 Flashwon 5 · lost 6
GPT-5.2won 5 · lost 6
Qwen3.8 27Bwon 5 · lost 6
Grok 4.5won 5 · lost 6
Gemini 3.7 Flashwon 5 · lost 6
Gemini 3.1 Prowon 5 · lost 6
Grok 4.20won 3 · lost 4
Gemini 3 Prowon 3 · lost 4
O4 Miniwon 3 · lost 5
GPT-5.6 Lunawon 4 · lost 7
GLM-5.3won 4 · lost 7
Claude Opus 4.8won 4 · lost 7
Gemini 3.8 Flashwon 4 · lost 7
Grok 4.6won 4 · lost 7
Grok 4.7won 4 · lost 7
Qwen3.8 Maxwon 4 · lost 7
GPT-5won 2 · lost 5
Grok 4won 2 · lost 5
DeepSeek V4 Prowon 3 · lost 8
Muse Spark 1.2won 3 · lost 8
Claude Opus 4.7won 3 · lost 8
GPT-5.4won 3 · lost 8
DeepSeek V4.1 Flashwon 3 · lost 8
Kimi K3won 3 · lost 8
Muse Spark 1.3won 3 · lost 8
O3won 2 · lost 6
GPT-5.6 Terrawon 2 · lost 9
GPT-6 Solwon 2 · lost 9
Claude Opus 5won 1 · lost 10
GPT-6 Astrawon 1 · lost 10
GPT-5.6 Solwon 0 · lost 11
Claude Opus 5.5won 0 · lost 11
GPT-5.5won 0 · lost 11
Claude Fable 5.1won 0 · lost 11
Claude Fable 5won 0 · lost 11

§ 3 · Sources

Where the numbers come from

1 publication, 8 figures. Every one links to the page it was read from.

LiveBench ↗8 measuresread 2026-09-26
LiveBench 76.1LiveBench · Reasoning 80.8LiveBench · Coding 82.1LiveBench · Agentic Coding 54.7LiveBench · Mathematics 95.3LiveBench · Data Analysis 74.6LiveBench · Language 85.9LiveBench · Instruction Following 59.5

Badge

PublicAI Index badge for Union Alpha[![PublicAI Index](https://publicai.io/model-index/badge?model=union-alpha)](https://publicai.io/model-index/m/union-alpha)