The LLM benchmark aggregator.

Inkling Small

Thinking Machines · 266B · open weights

Strongest in Cybersecurity (#34 of 47), weakest in Core abilities (#143 of 199). Above par in 9 of 19 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5.1 and Kimi K3 and ahead of O4 Mini and Claude 3.7 Sonnet.

§ 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.

Knowledge52.3−10.2#84/1971/3

Academic knowledge55.4−8.5#56/1231/1

Professional50.2−12.2#95/1941/2

Cybersecurity49.3−14.6#34/471/1

Finance54.4−8#58/1621/1

Medical49.9−16.6#87/1491/1

Legal45.3−20.6#118/1731/1

Agents51.7−14.7#104/2982/6

Knowledge work52.7−18#81/2082/2

Coding45.9−20.3#115/1791/6

Agentic coding45.7−21.2#111/1731/5

Human preference57.2−11.6#128/3541/1

Reasoning45.6−19.8#134/1952/5

Science55.2−8.3#57/1221/1

Reasoning50−19#67/1471/3

Mathematics35.5−28.3#139/1521/2

Core abilities45.3−20#143/1991/4

General intelligence43.2−26.2#134/1781/3

§ 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 Inkling Small, left for the other.

Mistral Small 3.1 24B Instructwon 13 · lost 0

NVIDIA Nemotron 3.5 Lightning 30B A3Bwon 15 · lost 0

GPT-4o miniwon 17 · lost 0

GPT-4.1 nanowon 17 · lost 1

Llama 4 Scout Instructwon 15 · lost 1

Gemini 2.0 Flashwon 14 · lost 2

Claude 3.5 Haikuwon 12 · lost 2

Llama 3.3 70B Instructwon 11 · lost 2

Gemini 1.5 Prowon 10 · lost 2

GPT-5.4 nanowon 15 · lost 3

Mistral Medium 3.5won 14 · lost 3

Command Awon 13 · lost 3

Gemma 4 31Bwon 11 · lost 3

Llama 4 Maverick Instructwon 14 · lost 4

Claude 3.5 Sonnetwon 10 · lost 3

GPT-5.4 miniwon 13 · lost 4

Ling 3.0 Flashwon 9 · lost 3

GPT-4owon 12 · lost 4

Gemini 3.5 Flash-Litewon 14 · lost 5

GPT OSS 20Bwon 11 · lost 4

MiMo V2.5won 11 · lost 4

GPT OSS 120Bwon 13 · lost 5

Gemini 2.5 Flash-Litewon 10 · lost 4

O3 Miniwon 11 · lost 5

Claude Haiku 4.5won 11 · lost 5

Mistral Medium 3.1won 8 · lost 4

GPT-5 nanowon 10 · lost 5

DeepSeek V3won 12 · lost 6

MiniMax M2.5won 10 · lost 6

Claude Sonnet 4won 11 · lost 7

Mistral Large 3won 11 · lost 7

MiniMax M2.7won 9 · lost 6

MiMo V2.5 Prowon 9 · lost 6

Gemini 3.1 Flash-Litewon 10 · lost 7

Qwen3.6 Pluswon 9 · lost 7

GPT-4.1 miniwon 10 · lost 8

GPT-5 miniwon 10 · lost 8

Qwen3.5 Flashwon 7 · lost 6

Claude 3.7 Sonnetwon 8 · lost 7

O4 Miniwon 9 · lost 8

Kimi K2 Thinkingwon 7 · lost 7

MiniMax M2.1won 7 · lost 7

O1won 7 · lost 7

GLM-4.7won 8 · lost 8

Grok 4.3won 8 · lost 8

NVIDIA Nemotron 3 Ultra 550B A55Bwon 9 · lost 9

DeepSeek R1won 8 · lost 9

Grok 4.1 Fastwon 7 · lost 8

Grok 3 Miniwon 7 · lost 9

Grok 4 Fastwon 6 · lost 8

MiniMax M3won 8 · lost 11

Inklingwon 8 · lost 11

Kimi K2won 7 · lost 10

GLM-4.6won 7 · lost 10

Qwen3 Maxwon 6 · lost 9

DeepSeek V3.2won 7 · lost 11

Grok 3won 6 · lost 10

Grok 4.20won 6 · lost 10

Step 5won 5 · lost 9

Gemini 2.5 Flashwon 6 · lost 11

GPT-6 Lunawon 6 · lost 11

Gemini 2.5 Prowon 6 · lost 12

Kimi K2.5won 6 · lost 12

Claude Sonnet 4.5won 6 · lost 12

GPT-5.1won 5 · lost 11

GLM-4.5won 4 · lost 9

GPT-4.1won 5 · lost 12

Qwen3 235B A22B Instructwon 5 · lost 12

Gemini 3.1 Prowon 5 · lost 12

Mistral Large 4won 4 · lost 11

Qwen3.8 27Bwon 5 · lost 14

Claude Haiku 5.5won 4 · lost 13

GPT-5won 4 · lost 14

Claude Opus 4won 3 · lost 13

GLM-5won 3 · lost 13

GLM-5.1won 3 · lost 13

Grok 4won 3 · lost 14

Qwen3.7 Maxwon 3 · lost 14

Claude Opus 4.1won 2 · lost 10

Gemini 3 Flashwon 2 · lost 12

Muse Sparkwon 2 · lost 13

Grok 4.7won 2 · lost 14

Kimi K2.6won 2 · lost 14

O3won 2 · lost 15

GPT-6 Astrawon 2 · lost 15

Claude Opus 4.5won 2 · lost 15

GPT-6.1 Solwon 2 · lost 15

Grok 4.5won 2 · lost 17

GLM-5.3 Flashwon 2 · lost 17

Gemini 3.8 Flashwon 2 · lost 17

GPT-6 Solwon 1 · lost 15

MiMo V2.6 Prowon 1 · lost 15

Claude Opus 4.6won 1 · lost 15

Claude Sonnet 5.5won 1 · lost 16

Gemini 3.5 Flashwon 1 · lost 16

Claude Opus 5.5won 1 · lost 16

Claude Sonnet 5won 1 · lost 17

Gemini 3.6 Flashwon 1 · lost 17

Gemini 3.7 Flashwon 1 · lost 17

DeepSeek V4 Flashwon 1 · lost 18

GPT-5.6 Lunawon 1 · lost 18

Qwen3.8 Maxwon 1 · lost 18

Gemini 4 Carbonwon 0 · lost 12

Muse Wonder Fullwon 0 · lost 13

Inkling 2 Smallwon 0 · lost 13

Kimi K3.1 Prowon 0 · lost 13

MiMo V2.6 Flashwon 0 · lost 15

Muse Spark 1.3won 0 · lost 15

Muse Wonderwon 0 · lost 15

Muse Spark 1.1won 0 · lost 16

Gemini 4 Argonwon 0 · lost 16

GPT-5.2won 0 · lost 17

Claude Sonnet 4.6won 0 · lost 17

DeepSeek V4.1 Flashwon 0 · lost 17

Gemini 3 Prowon 0 · lost 17

Muse Spark 1.2won 0 · lost 17

GLM-5.2won 0 · lost 18

GPT-5.4won 0 · lost 18

GPT-5.5won 0 · lost 18

Claude Opus 4.8won 0 · lost 18

Claude Opus 5won 0 · lost 18

Claude Opus 4.7won 0 · lost 18

Claude Fable 5won 0 · lost 18

Grok 4.6won 0 · lost 19

DeepSeek V4 Prowon 0 · lost 19

GPT-5.6 Terrawon 0 · lost 19

GLM-5.3won 0 · lost 19

GPT-5.6 Solwon 0 · lost 19

Kimi K3won 0 · lost 19

Claude Fable 5.1won 0 · lost 19

§ 3 · Sources

Where the numbers come from

6 publications, 21 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-11-10

LMArena Text 1405

Artificial Analysis ↗1 measureread 2026-11-10

Artificial Analysis Intelligence Index 13

GDPval-AA ↗1 measureread 2026-11-10

GDPval-AA 1133

AA-Briefcase ↗1 measureread 2026-11-10

AA-Briefcase 913

ARC-AGI-2 ↗1 measureread 2026-11-10

ARC-AGI-2 40.1%

Vals.ai ↗16 measuresread 2026-11-10

Vals · Legal Research Bench 25.48%Vals · LegalBench 82.99%Vals · Harvey Legal Agent Benchmark 2.92%Vals · Finance Agent 41.26%Vals · CorpFin 69.62%Vals · TaxEval 75.51%Vals · MortgageTax 62.28%Vals · MedCode 37.89%Vals · MedScribe 84.11%Vals · CyberBench 64.11%Vals · SWE-bench Verified 82.2% (Mini-SWE-agent)Vals · Vibe Code Bench 19.05% (OpenHands)Vals · Code Migration 13.68%Vals · GPQA Diamond 83.59%Vals · MMLU Pro 85.57%Vals · ProofBench 6%

Badge

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