Inkling 2 Mini

  • New
  • Open weights
  • Reasoning
  • Input: text · image
  • Released 2026-11
RankedEnough results for an index.
IndexAll categories, out of 100.57.7±4.2Median 50.7
CoverageIndex weight that has results.90%
SpeedOutput tokens per second.88/sMedian 76/s
First tokenSeconds to the first output token.0.7 sMedian 1.1 s
Input / 1MUS dollars per 1M input tokens.$0.10Median $0.60
Output / 1MUS dollars per 1M output tokens.$0.40Median $2.50
ContextMaximum tokens in one request.1.05MMedian 262K
EloLMArena rating and rank.—Median 1425

Details

From OpenRouter
API IDThe model ID on OpenRouter.
thinkingmachines/inkling-2-mini
Max outputMaximum output tokens per request.
66K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.02
Weights
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.8 counted60.568 / 221
CodingCode writing and repair.2 counted60.068 / 214
ReasoningLogic problems and puzzles.3 counted61.660 / 225
MultimodalTasks with images and text.8 counted57.149 / 88
KnowledgeFacts and expert knowledge.3 counted49.6136 / 266
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.None counted——
MathMath problems.1 counted47.4—

Results

25 counted
BenchmarkResultThe score from the source.PlacePlace among results on this benchmark.
LongBench v2Reasoning66.9%5 / 25
MCP AtlasAgentic82.7%12 / 53
CharXiv ReasoningMultimodal88.3%15 / 50
Graduate-Level Google-Proof Q&AKnowledge86.6%47 / 144
MathVisionMultimodal89.0%10 / 25
ARC-AGI-2 (semi-private)Reasoning49.3%55 / 137
ARC-AGI-1 (semi-private)Reasoning84.5%56 / 132
ScreenSpot ProMultimodal76.5%10 / 22
JobBenchAgentic49.6%17 / 37
DeepSearchQAAgentic85.3%12 / 25
BrowseCompAgentic79.0%31 / 63
GPQA Diamond · BenchLMKnowledge86.6%53 / 87
BabyVisionMultimodal53.4%8 / 12
ApexMath34.4%7 / 10
DeepSWEAgentic48.2%56 / 64
ZeroBenchMultimodal17.0%10 / 11
FrontierCode 2.0Coding15.1%31 / 32
Chartography without toolsMultimodal22.1%12 / 12
Terminal-Bench-Science 0.1Agentic1.4%15 / 15
ARC-AGI-2 (public eval)Reasoning51.2%48 / 127
ARC-AGI-1 (public eval)Reasoning90.5%49 / 116
OmniDocBench v1.6Multimodal92.8%—
BabyVision with PythonMultimodal88.8%—
PerceptionBench (Internal)Multimodal57.5%—
ZeroBench_main with PythonMultimodal34.0%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseARC Prize FoundationTerms restrict reuse

Questions

How much does Inkling 2 Mini cost?
$0.10 per 1M input tokens, $0.40 per 1M output tokens.
How good is Inkling 2 Mini?
Its index is 57.7 out of 100. That is place 88 of 282 ranked models.
What is Inkling 2 Mini best at?
Reasoning: 61.6 out of 100, place 60 of 225.
How fast is Inkling 2 Mini?
88 output tokens per second, 0.7 s to the first token.
What is the context window of Inkling 2 Mini?
1.05M tokens.
Is Inkling 2 Mini open source?
Yes. Its weights are published.
When was Inkling 2 Mini released?
November 2026.

Badge

Inkling 2 Mini on Inference360: place 88 of 282, index 57.7

Compare with

Beam 1.0 501B-A23B NVFP457.6Kimi K2.657.4Apodex 1.158.1Beam 1.058.1dots3-note Preview62.3

Lower price

dots3-note Preview62.3 · FreeApodex 1.1 Mini54.9 · FreeGLM-5.3-Flash62.4 · $0.078/1M

More from Thinking Machines Lab

All models
Inkling 2 Small65.5Inkling-Small54.5Inkling54.1