The LLM benchmark aggregator.

Command A (03-2025)

Cohere · 111B · open weights

Strongest in Tool use (#26 of 81), weakest in Fairness (#287 of 296). Above par in 8 of 24 scopes. Among the models it meets almost everywhere, it finishes behind GPT-5.4 and Claude Sonnet 5 and ahead of GPT-4o mini and Claude 3.5 Haiku.

§ 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.

Agents52.6−13.3#74/2891/6

Tool use56.3−17.7#26/811/1

Professional47.5−15.6#122/1881/2

Legal54.6−11.3#45/1641/1

Medical48.4−18.1#98/1491/1

Finance39−23.5#142/1611/1

Knowledge43.7−19#160/1911/3

Academic knowledge34.9−29#110/1231/1

Coding39.7−26.7#166/1761/6

Agentic coding38.7−28.3#162/1691/5

Human preference52.3−16.3#168/3501/1

Reasoning36.7−28.2#175/1901/5

Science30.8−32.7#110/1221/1

Mathematics35.9−28.2#134/1501/2

Core abilities41.9−23.3#197/2251/4

General intelligence38.3−30.5#192/2141/3

Safety46.1−15.3#269/3312/3

Factual grounding51.8−19#50/1021/1

Toxicity avoidance53.1−5.9#149/2711/1

Secure code49.5−16.8#159/2731/1

Jailbreak resistance51.7−13.3#161/2711/1

Harm refusal41.8−19.7#272/2961/2

Fairness36.9−33.1#287/2961/2

§ 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 Command A (03-2025), left for the other.

Jamba 1.6 Miniwon 12 · lost 3

Command R+ (08-2024)won 12 · lost 4

Jamba 1.6 Largewon 11 · lost 4

Mistral Largewon 12 · lost 7

Gemini 1.5 Flashwon 10 · lost 6

Llama 4 Scout Instructwon 13 · lost 9

Mistral Smallwon 11 · lost 8

Claude 3.5 Haikuwon 11 · lost 8

GPT-4o miniwon 12 · lost 9

Gemini 1.5 Prowon 8 · lost 9

Mistral Medium 3.5won 7 · lost 9

GPT-4.1 nanowon 10 · lost 13

NVIDIA-Nemotron-3-Ultra-550B-A55B (NVFP4)won 10 · lost 13

Llama 3.3 70B Instructwon 8 · lost 11

Claude Haiku 4.5won 10 · lost 14

Grok 4.1 Fastwon 8 · lost 14

GPT OSS 20Bwon 7 · lost 13

GPT-4owon 7 · lost 14

Gemma 4 31Bwon 7 · lost 14

Llama 4 Maverick Instructwon 8 · lost 16

DeepSeek R1won 7 · lost 15

GLM-4.7won 5 · lost 11

GPT OSS 120Bwon 7 · lost 16

Ling 3.0 Flashwon 5 · lost 12

Grok 4.3won 6 · lost 15

GPT-5.4 nanowon 5 · lost 13

GPT-5 nanowon 6 · lost 16

Gemma 4 E4Bwon 4 · lost 11

DeepSeek V3.1won 4 · lost 11

Gemini 2.5 Flash-Litewon 5 · lost 14

Gemini 3.5 Flash-Litewon 4 · lost 12

GLM-5.1won 5 · lost 15

MiniMax M2.7won 5 · lost 16

Qwen3.7 Maxwon 5 · lost 16

GPT-5.4 miniwon 4 · lost 13

Grok 3won 5 · lost 17

DeepSeek V4 Flashwon 5 · lost 17

Kimi K2.6won 5 · lost 17

Gemini 3.1 Flash-Litewon 4 · lost 14

DeepSeek V3won 5 · lost 18

MiniMax M3won 5 · lost 18

GLM-4.5won 4 · lost 15

MiMo V2.6 Flashwon 4 · lost 15

Gemma 3 27B ITwon 3 · lost 12

Qwen3.5 Flashwon 3 · lost 12

Qwen3.6 Pluswon 3 · lost 12

Grok 4.20 (Reasoning, 0309)won 3 · lost 12

O3 Miniwon 4 · lost 16

GPT-6 Lunawon 4 · lost 16

MiMo V2.5 Prowon 4 · lost 16

Gemini 2.0 Flashwon 4 · lost 17

Gemini 3.6 Flashwon 4 · lost 17

NVIDIA Nemotron 3 Super 120B A12Bwon 3 · lost 13

Inkling Smallwon 3 · lost 13

Qwen3.8 27Bwon 3 · lost 14

Claude Sonnet 4won 4 · lost 19

Grok 4won 4 · lost 19

Mistral Large 3won 3 · lost 15

Gemini 3.1 Prowon 3 · lost 15

O4 Miniwon 4 · lost 20

Kimi K2won 4 · lost 20

GLM-4.6won 4 · lost 20

Mistral Large 4won 3 · lost 16

Claude 3.7 Sonnetwon 3 · lost 17

GPT-5.1won 3 · lost 18

GPT-6 Solwon 3 · lost 18

Gemini 3.5 Flashwon 3 · lost 18

MiniMax M2.1won 2 · lost 13

Qwen3 Maxwon 2 · lost 13

Gemini 3 Flashwon 2 · lost 13

GPT-4.1 miniwon 3 · lost 20

Qwen3 235B A22B Instructwon 3 · lost 20

O3won 3 · lost 20

Inklingwon 3 · lost 20

Gemini 2.5 Prowon 3 · lost 20

Kimi K2.5won 3 · lost 20

GPT-5.6 Lunawon 3 · lost 20

GPT-5 miniwon 3 · lost 21

Gemini 2.5 Flashwon 3 · lost 21

DeepSeek V3.2won 3 · lost 21

Grok 4 Fastwon 2 · lost 15

GLM-5won 2 · lost 15

Claude Opus 4.6won 2 · lost 18

Claude Sonnet 4.6won 2 · lost 19

Gemini 3 Prowon 2 · lost 19

Claude Opus 4won 2 · lost 20

Grok 4.5won 2 · lost 20

GPT-5won 2 · lost 21

DeepSeek-V4-Pro (High, 20260813)won 2 · lost 21

GLM-5.2won 2 · lost 21

GPT-5.5won 2 · lost 21

GPT-5.6 Solwon 2 · lost 21

Claude Sonnet 4.5 (High, 32K, 20250929)won 2 · lost 22

Muse Spark 1.2won 1 · lost 14

Gemini 4 Argonwon 1 · lost 14

GPT-6 Astrawon 1 · lost 15

GLM-5.3won 1 · lost 15

MiniMax M2.5won 1 · lost 16

Grok 3 Miniwon 1 · lost 17

Claude 3.5 Sonnetwon 1 · lost 17

Kimi K2 Thinkingwon 1 · lost 18

O1won 1 · lost 18

Claude Opus 4.5 (High, 32K, 20251101)won 1 · lost 18

Claude Opus 5.5won 1 · lost 19

Muse Spark 1.1won 1 · lost 20

GLM-5.3 Flashwon 1 · lost 21

Qwen3.8 Maxwon 1 · lost 21

Claude Opus 5won 1 · lost 21

Gemini 3.8 Flashwon 1 · lost 21

GPT-5.6 Terrawon 1 · lost 22

Claude Opus 4.7won 1 · lost 22

GPT-4.1won 1 · lost 23

GPT-5.2won 1 · lost 23

DeepSeek V4.1 Flashwon 0 · lost 15

Kimi K3won 0 · lost 16

Gemini 3.7 Flashwon 0 · lost 16

Claude Fable 5.1won 0 · lost 16

Claude Fable 5won 0 · lost 16

Grok 4.7won 0 · lost 19

MiMo V2.6 Prowon 0 · lost 19

Grok 4.6won 0 · lost 22

Claude Opus 4.8won 0 · lost 22

Claude Sonnet 5won 0 · lost 23

GPT-5.4won 0 · lost 23

§ 3 · Sources

Where the numbers come from

6 publications, 19 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-10-21

LMArena Text 1354

BFCL ↗1 measureread 2026-10-21

BFCL v4 46.49%

Kagi LLM Benchmark ↗1 measureread 2026-10-21

Kagi LLM Benchmark 28.8%

Vals.ai ↗9 measuresread 2026-10-21

Vals · CaseLaw 64.52%Vals · LegalBench 79.7%Vals · CorpFin 45.96%Vals · TaxEval 61.37%Vals · MedQA 80.55%Vals · SWE-bench Verified 7.8% (Mini-SWE-agent)Vals · GPQA Diamond 48.48%Vals · MMLU Pro 69.17%Vals · AIME 13.33%

Enkrypt AI Safety Leaderboard ↗6 measuresread 2026-10-21

Enkrypt · Jailbreak risk 11.3%Enkrypt · Harmful content risk 51.1%Enkrypt · CBRN risk 27.3%Enkrypt · Toxicity risk 4%Enkrypt · Bias risk 98.2%Enkrypt · Insecure code risk 34.2%

Vectara Hallucination Leaderboard ↗1 measureread 2026-10-21

Vectara · Factual consistency 90.7%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=command-a-03-2025)](https://publicai.io/model-index/m/command-a-03-2025)