The LLM benchmark aggregator.

Mercury 2

Inception AI

Strongest in Factual grounding (#81 of 101), weakest in Safety (#233 of 336). Above par in 2 of 6 scopes. Among the models it meets almost everywhere, it finishes behind Claude Opus 4.7 and Gemini 3.1 Pro and ahead of Gemma 3 4B It and Ministral 8B.

§ 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.

Human preference51.5−16#174/3411/1

Agents42.7−25.4#213/2672/5

Knowledge work40.5−33.1#130/1772/2

Safety48.2−13.3#233/3361/3

Factual grounding44.1−26.5#81/1011/1

§ 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 Mercury 2, left for the other.

Ministral 3 3Bwon 4 · lost 0

Phi-4 Mini Instructwon 4 · lost 0

Ministral 3 8Bwon 4 · lost 0

Ministral 3 14Bwon 4 · lost 0

Granite 4.2 3Bwon 4 · lost 0

Llama 3.3 70B Instructwon 4 · lost 0

Llama 3.1 8B Instructwon 5 · lost 0

GPT-4won 4 · lost 1

GPT-4o Miniwon 4 · lost 1

GPT-4.1 Nanowon 4 · lost 1

Llama 3.2 1B Instructwon 3 · lost 1

Command R Pluswon 3 · lost 1

Llama 4 Scout Instructwon 4 · lost 2

Llama 4 Maverick Instructwon 4 · lost 2

Gemma 3 12B Itwon 4 · lost 2

Mistral Large 3won 4 · lost 2

Ministral 8Bwon 3 · lost 2

Gemma 3 4B Itwon 3 · lost 2

NVIDIA Nemotron 3 Nano 30B A3Bwon 2 · lost 2

Qwen3 8Bwon 2 · lost 2

Llama 3.2 3B Instructwon 2 · lost 2

Snowflake Arctic Instructwon 2 · lost 2

Granite 3.1 8B Instructwon 2 · lost 2

C4ai Aya Expanse 8Bwon 2 · lost 2

Nova Litewon 2 · lost 2

C4ai Aya Expanse 32Bwon 2 · lost 2

Granite 4.2 8Bwon 2 · lost 2

GPT-4owon 2 · lost 2

Granite 4.2 30Bwon 2 · lost 2

Qwen3 235B A22B Thinkingwon 2 · lost 2

Grok 4 Fastwon 2 · lost 2

Gemini 3 Flashwon 2 · lost 2

DeepSeek V3won 3 · lost 3

DeepSeek R1won 3 · lost 3

Nova Microwon 2 · lost 3

Phi 4won 2 · lost 3

Nova Prowon 2 · lost 3

Mistral Largewon 2 · lost 3

GPT OSS 20Bwon 2 · lost 3

Claude 3.5 Haikuwon 2 · lost 3

GPT-5 Nanowon 2 · lost 3

O3 Miniwon 2 · lost 3

Mistral Mediumwon 2 · lost 3

Grok 4.1 Fastwon 2 · lost 3

Nova 2 Litewon 2 · lost 4

Gemma 3 27B Itwon 2 · lost 4

Trinity Largewon 2 · lost 4

Gemini 3.1 Flash Litewon 2 · lost 4

Qwen3 14Bwon 1 · lost 3

Qwen3 30B A3B Instructwon 1 · lost 3

Grok 3won 1 · lost 3

Grok 4won 1 · lost 3

Command Awon 1 · lost 4

NVIDIA Nemotron 3 Super 120B A12Bwon 1 · lost 4

O4 Miniwon 1 · lost 4

Kimi K2won 1 · lost 4

DeepSeek V4 Flashwon 1 · lost 4

MiniMax M3won 1 · lost 4

Grok 4.3won 1 · lost 4

GLM-5.2won 1 · lost 4

Gemini 3.5 Flashwon 1 · lost 4

Gemini 3.6 Flashwon 1 · lost 4

Gemini 3 Prowon 1 · lost 4

Qwen3 32Bwon 1 · lost 5

GPT OSS 120Bwon 1 · lost 5

Qwen3 Next 80B A3B Thinkingwon 1 · lost 5

GPT-5 Miniwon 1 · lost 5

GLM-4.6won 1 · lost 5

GPT-5won 1 · lost 5

Kimi K2.5won 1 · lost 5

GPT-5.6 Solwon 1 · lost 5

GLM-4.7 Flashwon 0 · lost 4

GLM-4.5 Airwon 0 · lost 4

Mimo V2 Flashwon 0 · lost 4

Claude 3.7 Sonnetwon 0 · lost 4

Qwen3.5 Flashwon 0 · lost 4

Inkling Smallwon 0 · lost 4

Qwen3.5 27Bwon 0 · lost 4

DeepSeek V3.1won 0 · lost 4

DeepSeek V3.1 Terminuswon 0 · lost 4

Muse Glimmerwon 0 · lost 4

Mistral Medium 3.5won 0 · lost 4

Claude Opus 4won 0 · lost 4

O3won 0 · lost 4

Mimo V2.5won 0 · lost 4

Qwen3.8 27Bwon 0 · lost 4

GLM-4.7won 0 · lost 4

Qwen3.6 Pluswon 0 · lost 4

Claude Opus 4.1won 0 · lost 4

Gemini 3.5 Flash Litewon 0 · lost 4

Qwen3.7 Pluswon 0 · lost 4

GLM-5won 0 · lost 4

GPT-5.5 Instantwon 0 · lost 4

DeepSeek V4.1 Flashwon 0 · lost 4

GLM-5.3won 0 · lost 4

Gemini 3.7 Flashwon 0 · lost 4

Kimi K3won 0 · lost 4

Muse Sparkwon 0 · lost 4

Muse Spark 1.3won 0 · lost 4

Muse Spark 1.2won 0 · lost 4

Claude Fable 5.1won 0 · lost 4

Claude Fable 5won 0 · lost 4

Claude Opus 4.6won 0 · lost 4

Mistral Smallwon 0 · lost 5

Gemini 2.5 Flash Litewon 0 · lost 5

GPT-4.1 Miniwon 0 · lost 5

Gemini 2.5 Flashwon 0 · lost 5

GPT-4.1won 0 · lost 5

MiniMax M2.7won 0 · lost 5

Qwen3 235B A22B Instructwon 0 · lost 5

GPT-5.2won 0 · lost 5

Grok 4.7won 0 · lost 5

GPT-6 Lunawon 0 · lost 5

Inklingwon 0 · lost 5

Qwen3.5 397B A17Bwon 0 · lost 5

Grok 4.6won 0 · lost 5

GPT-5.6 Lunawon 0 · lost 5

Hy3won 0 · lost 5

Claude Sonnet 5won 0 · lost 5

GLM-5.1won 0 · lost 5

Grok 4.5won 0 · lost 5

GPT-5.6 Terrawon 0 · lost 5

Mimo V2.5 Prowon 0 · lost 5

GLM-5.3 Flashwon 0 · lost 5

Claude Opus 4.5won 0 · lost 5

Qwen3.7 Maxwon 0 · lost 5

Qwen3.8 Maxwon 0 · lost 5

Mimo V2.6 Prowon 0 · lost 5

Claude Opus 4.8won 0 · lost 5

Claude Opus 5won 0 · lost 5

Muse Spark 1.1won 0 · lost 5

Gemini 3.8 Flashwon 0 · lost 5

Claude Opus 5.5won 0 · lost 5

Qwen3.5 35B A3Bwon 0 · lost 6

GPT-5.4 Nanowon 0 · lost 6

Claude Sonnet 4won 0 · lost 6

Claude Haiku 4.5won 0 · lost 6

Qwen3.5 122B A10Bwon 0 · lost 6

DeepSeek V3.2won 0 · lost 6

Gemma 4 26B A4Bwon 0 · lost 6

Gemini 2.5 Prowon 0 · lost 6

GPT-5.4 Miniwon 0 · lost 6

Gemma 4 31Bwon 0 · lost 6

GPT-5.1won 0 · lost 6

GPT-6 Solwon 0 · lost 6

Claude Sonnet 4.5won 0 · lost 6

Kimi K2.6won 0 · lost 6

DeepSeek V4 Prowon 0 · lost 6

Claude Sonnet 4.6won 0 · lost 6

GPT-5.4won 0 · lost 6

GPT-6 Astrawon 0 · lost 6

GPT-5.5won 0 · lost 6

Gemini 3.1 Prowon 0 · lost 6

Claude Opus 4.7won 0 · lost 6

§ 3 · Sources

Where the numbers come from

4 publications, 4 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-09-27

LMArena Text 1343

GDPval-AA ↗1 measureread 2026-09-27

GDPval-AA 481

AA-Briefcase ↗1 measureread 2026-09-27

AA-Briefcase 264

Vectara Hallucination Leaderboard ↗1 measureread 2026-09-27

Vectara · Factual consistency 87.7%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=mercury-2)](https://publicai.io/model-index/m/mercury-2)