The LLM benchmark aggregator.

K2 Horizon 7B

Institute of Foundation Models · 7B

Strongest in Knowledge work (#50 of 178), weakest in General intelligence (#122 of 204). Above par in 6 of 13 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and GPT-6 Astra and ahead of Gemini 3.1 Pro and GPT-5 Mini.

§ 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.

Agents54−14.1#74/2682/5

Knowledge work55.2−18.4#50/1782/2

Tool use54.7−19.3—/81✱0/1

Web research41.4−22.8—/4✱0/1

Core abilities48−19.4#118/2041/3

General intelligence47.1−23#122/2041/3

Long context49.5−15.1—/0✱✱

Reasoning54.2−12.7—/178✱0/4

Mathematics56.3−9.4—/140✱0/2

Expert reasoning60.9leads—/0✱✱

Coding46.4−22.7—/165✱0/5

Agentic coding46.6−22.1—/157✱0/4

Code generation46.8−23.2—/77✱0/2

§ 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 K2 Horizon 7B, left for the other.

Mistral Medium 3.5won 8 · lost 0

Mercury 2.5won 8 · lost 0

Mistral Large 3won 8 · lost 0

NVIDIA Nemotron 3 Nano 4Bwon 9 · lost 0

Qwen3.6 27Bwon 9 · lost 0

Grok 4.3won 9 · lost 0

GPT-5.4 Miniwon 9 · lost 0

GPT-5.4 Nanowon 9 · lost 0

Nemotron 3 Ultra 550B A55Bwon 9 · lost 0

GPT-4o Miniwon 9 · lost 0

GPT-4.1 Nanowon 9 · lost 0

Gemini 3.5 Flash Litewon 8 · lost 1

Qwen3.6 Pluswon 8 · lost 1

Grok Build 0.1won 8 · lost 1

G9v3 3Bwon 7 · lost 1

Command Awon 7 · lost 1

GPT OSS 120Bwon 7 · lost 1

Inklingwon 11 · lost 2

Granite 4.2 8Bwon 10 · lost 2

Granite 4.2 3Bwon 9 · lost 2

K2 Horizon 3.7Bwon 9 · lost 2

Kimi K2.7 Codewon 7 · lost 2

K2 Horizon 0.9Bwon 7 · lost 2

Claude Haiku 4.5won 7 · lost 2

Llama 4 Maverick Instructwon 7 · lost 2

Qwen3.7 Maxwon 7 · lost 2

Qwen3.5 4Bwon 6 · lost 2

Gemini 3.1 Flash Litewon 6 · lost 2

Qwen3.6 35B A3Bwon 8 · lost 3

Qwen3.5 9Bwon 8 · lost 3

Granite 4.2 30Bwon 7 · lost 3

Nemotron 3 Superwon 7 · lost 3

Gemma 4 12Bwon 7 · lost 3

MiniMax M3won 9 · lost 4

GLM-4.6won 6 · lost 3

G9v3 39A5Bwon 6 · lost 3

K2 Horizon 32Bwon 6 · lost 3

DeepSeek V3won 6 · lost 3

Nemotron 3 Ultrawon 8 · lost 4

Kimi K2.6won 5 · lost 4

Gemini 2.5 Prowon 5 · lost 4

GPT-5 Miniwon 5 · lost 4

Gemini 3.1 Prowon 5 · lost 4

Mimo V2.5 Prowon 4 · lost 4

Kimi K2.5won 4 · lost 4

Gemma 4 31Bwon 5 · lost 6

Claude Sonnet 4.6won 4 · lost 5

GPT-4.1 Miniwon 4 · lost 5

Claude Sonnet 4.5won 4 · lost 5

Muse Glimmer-30Bwon 4 · lost 5

DeepSeek R1won 4 · lost 5

Claude Sonnet 4won 4 · lost 5

DeepSeek V4 Flashwon 4 · lost 6

Gemini 3.5 Flashwon 4 · lost 6

GPT-5won 3 · lost 5

GPT-6 Lunawon 3 · lost 6

GLM-5.3 Flashwon 3 · lost 6

Muse Spark 1.1won 3 · lost 6

DeepSeek V3.2won 3 · lost 6

Claude Opus 4.5won 3 · lost 6

Qwen3.8 Flash Nextwon 3 · lost 6

O3won 3 · lost 6

DeepSeek V4 Flash Visionwon 3 · lost 6

GLM-5.2won 4 · lost 8

Qwen3.8 27Bwon 4 · lost 8

GPT-5.6 Lunawon 4 · lost 9

GLM-5.3won 3 · lost 7

Grok 4.5won 3 · lost 7

K2 Horizon MoVA 36B A4Bwon 3 · lost 8

Kimi K2won 2 · lost 6

GPT-4.1won 2 · lost 6

Qwen3 235B A22B Instructwon 2 · lost 6

Mimo V2.6 Prowon 2 · lost 6

Gemini 3.6 Flashwon 2 · lost 7

Gemini 2.5 Flashwon 2 · lost 7

O4 Miniwon 2 · lost 7

Gemini 3.8 Flashwon 2 · lost 7

Grok 4.7won 2 · lost 7

Qwen3.8 Maxwon 2 · lost 7

DeepSeek V4 Prowon 2 · lost 8

DeepSeek V4.1 Flashwon 2 · lost 8

Kimi K3won 2 · lost 8

K2 Horizon 375B A23Bwon 2 · lost 10

Muse Spark 1.2won 1 · lost 8

Gemini 3.7 Flashwon 1 · lost 8

Grok 4.6won 1 · lost 8

Claude Opus 5won 1 · lost 9

GPT-5.6 Solwon 1 · lost 10

GPT-5.6 Terrawon 1 · lost 11

Claude Sonnet 5won 1 · lost 12

Grok 4won 0 · lost 8

Gemini 3 Prowon 0 · lost 8

GPT-5.2won 0 · lost 9

Claude Opus 4.8won 0 · lost 9

Claude Opus 4.7won 0 · lost 9

GPT-5.4won 0 · lost 9

GPT-6 Solwon 0 · lost 9

Claude Opus 5.5won 0 · lost 9

Muse Spark 1.3won 0 · lost 9

GPT-5.5won 0 · lost 9

Claude Fable 5.1won 0 · lost 9

GPT-6 Astrawon 0 · lost 9

Claude Fable 5won 0 · lost 10

§ 3 · Sources

Where the numbers come from

5 publications, 12 figures. Every one links to the page it was read from.

Artificial Analysis ↗1 measureread 2026-09-26

Artificial Analysis Intelligence Index 21

GDPval-AA ↗1 measureread 2026-09-26

GDPval-AA 1011

AA-Briefcase ↗1 measureread 2026-09-26

AA-Briefcase 1018

✱✱ IFM — K2 Horizon launch ↗the model’s own publisher8 measuresread 2026-09-26

tau3-Banking 25.8%BrowseComp 59%Terminal-Bench 2.1 39.1%SciCode 31.6%AA-LCR 68%SWE-bench Verified 70.6%HMMT Feb 2026 73.3%HLE 18.6%

✱ CellCog — K2 Horizon analysis ↗1 measureread 2026-09-26

Terminal-Bench 2.1 39.1%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=k2-horizon-7b)](https://publicai.io/model-index/m/k2-horizon-7b)