Inference
EN
Thinking Machines Lab

Inkling 2 Small

Inkling 2 Small is a reasoning model from Thinking Machines Lab in the Inkling family. 33 benchmarks count toward its index, in 6 categories.

  • New
  • Closed weights
  • Reasoning
  • Input: text
  • Released 2026-11
RankedEnough results for an index.
IndexAll categories, out of 100.66.4±3.8Median 51.0
CoverageIndex weight that has results.90%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.30Median $0.60
Output / 1MUS dollars per 1M output tokens.$1.00Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.1403±12 #153Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.14 counted67.635 / 218
CodingCode writing and repair.4 counted66.633 / 209
ReasoningLogic problems and puzzles.5 counted68.427 / 227
MultimodalTasks with images and text.6 counted63.630 / 84
KnowledgeFacts and expert knowledge.3 counted61.558 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.None counted——
MathMath problems.1 counted60.4—

Results

33 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
LongBench v2Reasoning70.6%1 / 25
JobBenchAgentic70.5%2 / 36
Artificial Analysis Long Context ReasoningReasoning83.3%13 / 217
Harvard-MIT Mathematics Tournament February 2026Math95.8%2 / 27
Graduate-Level Google-Proof Q&AKnowledge93.0%11 / 141
ARC-AGI-1 (semi-private)Reasoning96.5%11 / 126
MCP AtlasAgentic85.0%5 / 52
Software Engineering Benchmark VerifiedCoding82.2%11 / 114
MathVisionMultimodal92.8%3 / 24
Artificial Analysis Humanity's Last ExamKnowledge44.1%29 / 222
Massive Multi-discipline Multimodal Understanding ProMultimodal81.6%7 / 53
Artificial Analysis MMMU-ProMultimodal81.6%15 / 113
GPQA Diamond · BenchLMKnowledge93.0%12 / 84
BrowseCompAgentic90.2%9 / 63
ARC-AGI-2 (semi-private)Reasoning81.7%20 / 131
Humanity's Last ExamKnowledge53.2%16 / 77
Terminal-Bench 2.1 · BenchLMAgentic85.1%19 / 86
Vibe Code Bench v1.1Coding81.5% ±3.528 / 123
Agent Arena steerabilityAgentic9.313 / 54
DeepSWEAgentic72.1%17 / 59
CharXiv Reasoning without toolsMultimodal85.6%5 / 17
CharXiv ReasoningMultimodal87.7%15 / 49
Artificial Analysis AnalystAgentAgentic51.2%8 / 26
Humanity's Last Exam without toolsKnowledge44.1%17 / 53
ScreenSpot ProMultimodal83.8%7 / 20
Artificial Analysis SciCodeCoding52.3%41 / 114
DeepSearchQAAgentic90.9%9 / 24
Agent Arena task outcomeAgentic3.323 / 54
Artificial Analysis Omniscience AccuracyKnowledge33.8%71 / 164
BabyVisionMultimodal68.6%5 / 11
Artificial Analysis GDP.pdfAgentic25.2%11 / 22
Artificial Analysis Tau3-BankingAgentic40.4%13 / 22
Medical Long Context Reasoning (MLCR-AA)Reasoning20.6%14 / 22
FrontierCode 2.0Coding22.6%18 / 27
Terminal-Bench 4.0Agentic12.6% ±3.141 / 61
Artificial Analysis Terminal-Bench v2.1Agentic81.3%17 / 21
Artificial Analysis Terminal-Bench v4.0Agentic13.1%19 / 22
Artificial Analysis AutomationBenchAgentic45.6%20 / 23
Tax Agent BenchFinance38.9% ±3.311 / 82
ARC-AGI-2 (public eval)Reasoning84.2%20 / 122
HLAB-Vals-1.1Legal21.7% ±3.36 / 35
ARC-AGI-1 (public eval)Reasoning97.4%20 / 116
Legal Research BenchLegal46.6% ±3.518 / 90
SAGEEducation88.6% ±1.88 / 29
ProgramBenchCoding0.5% ±0.519 / 65
CyberBench v1.1Other72.7% ±5.716 / 48
BioMysteryBenchScience70.0% ±1.314 / 30
MysteryMechanismScience25.2% ±2.917 / 30
SRE BenchOther1.9% ±0.825 / 36
Terminal-Bench-ScienceScience2.9% ±2.033 / 44
BabyVision with PythonMultimodal93.7%—
Chartography with image and code toolsMultimodal80.4%—
PerceptionBench (Internal)Multimodal64.7%—
ApexMath63.5%—
Artificial Analysis EnterpriseOps-GymAgentic46.8%—
SWE-MarathonCoding38.8%—
Chartography without toolsMultimodal36.7%—
PostTrainBench v1.2Coding31.9%—

Sources

BenchLMCC BY-NC 4.0ARC Prize FoundationTerms restrict reuseVals AINo licence statedLMArenaCC BY 4.0

History

5 releases
Index66.3 → 66.4
11 Nov 202612 Nov 2026
Output / 1M$1.00 → $1.00
11 Nov 202612 Nov 2026

Questions

How much does Inkling 2 Small cost?
$0.30 per 1M input tokens, $1.00 per 1M output tokens. A free tier is also available.
How good is Inkling 2 Small?
Its index is 66.4 out of 100. That is place 35 of 283 ranked models.
What is Inkling 2 Small best at?
Reasoning: 68.4 out of 100, place 27 of 227.
What is the context window of Inkling 2 Small?
1M tokens.
Is Inkling 2 Small open source?
No. Its weights are not published.
When was Inkling 2 Small released?
November 2026.

Badge

Inkling 2 Small on Inference360: place 35 of 283, index 66.4
HTML
<a href="https://inference360.ai/models/inkling-2-small"><img src="https://inference360.ai/badges/inkling-2-small.svg" alt="Inkling 2 Small on Inference360: place 35 of 283, index 66.4" height="20"></a>
Markdown
[![Inkling 2 Small on Inference360: place 35 of 283, index 66.4](https://inference360.ai/badges/inkling-2-small.svg)](https://inference360.ai/models/inkling-2-small)

Compare with

Sakana Fugu-Ultra66.3Hy4 preview66.4Claude Haiku 5.566.3Kimi K366.9dots3-note Preview63.4

Lower price

dots3-note Preview63.4 · FreeLing 3.1 Flash67.2 · $0.135/1MMiMo-V2.6-Flash64.3 · $0.28/1M

More from Thinking Machines Lab

All models
Inkling-Small55.1Inkling54.8