Thinkingmachines

Inkling Small

Generally availableOpen weights1M contextReasoning

Summary

Inkling Small is an open-weight language model from Thinkingmachines, released on 30 Jul 2026. It accepts text, images and PDFs and generates text, with a 1M-token context window and up to 1M output tokens per response. As of 18 Nov 2026, the lowest listed API price is $0.45 per million input tokens and $1.20 per million output tokens (DeepInfra), across 4 providers we track.

Multimodal reasoning model for visual analysis, planning, and tool use

Last verified 18 Nov 2026. Every figure on this page links to its source.

Specifications

DeveloperThinkingmachines
Release date30 Jul 2026 [models.dev]
StatusGenerally available [models.dev]
WeightsOpen weights [Hugging Face]
Licenseapache-2.0 [Hugging Face]
Parameters266B [Hugging Face]
Context window1M (1,000,000 tokens) [models.dev]
Max output1M tokens [models.dev]
InputText, images and PDFs [models.dev]
OutputText [models.dev]
Reasoning modeYes [models.dev]
Tool callingYes [models.dev]
Structured outputYes [models.dev]
Hugging Facethinkingmachines/Inkling-Small [Hugging Face]

Inkling Small API pricing by provider

USD per million tokens. Sorted by input price.

ProviderInputOutputCached inputContextSource
DeepInfra
thinkingmachines/Inkling-Small
$0.45$1.20$0.10524Kmodels.dev, 18 Nov 2026
OpenRouter
thinkingmachines/inkling-small
$0.45$1.20$0.10524KOpenRouter, 18 Nov 2026
Vercel AI Gateway
thinkingmachines/inkling-small
$0.45$1.20$0.101Mmodels.dev, 18 Nov 2026
Hugging Face Inference Providers
thinkingmachines/Inkling-Small
$0.50$1.20–524Kmodels.dev, 18 Nov 2026

Running Inkling Small locally

Estimated memory, from the parameter count and the architecture in config.json on Hugging Face. Includes about 0.8 GB of runtime overhead.

QuantizationWeightsTotal, 8K contextTotal, 32K context
Q4_K_M (4-bit)161 GB163 GB168 GB
Q8_0 (8-bit)283 GB285 GB289 GB
FP16 / BF16 (16-bit)532 GB534 GB538 GB

At 4-bit with an 8K conversation, Inkling Small needs about 163 GB, more than any single consumer graphics card; it needs a unified-memory machine, a multi-GPU setup or a split between graphics card and system RAM. Check your own hardware, or see Best computer for local AI on our sister site AIPCs.

Frequently asked questions

What is the context window of Inkling Small?

Inkling Small has a context window of 1,000,000 tokens (1M) and can generate up to 1,000,000 tokens in a single response, according to models.dev.

How much does the Inkling Small API cost?

Inkling Small is listed from $0.45 per million input tokens and $1.20 per million output tokens (DeepInfra). Prices last verified 18 Nov 2026.

Is Inkling Small open source?

Inkling Small is an open-weight model: its weights are published on Hugging Face as thinkingmachines/Inkling-Small under the apache-2.0 license. Check the license terms before commercial use.

When was Inkling Small released?

Thinkingmachines released Inkling Small on 30 Jul 2026, according to models.dev.

Which providers offer Inkling Small?

We track Inkling Small on 4 providers: DeepInfra, OpenRouter, Vercel AI Gateway, Hugging Face Inference Providers.

How much VRAM does Inkling Small need?

About 163 GB at 4-bit (Q4_K_M) with an 8K-token conversation, and about 285 GB at 8-bit. That is more than any single consumer graphics card holds. These are estimates from the model's published parameter count and architecture.

Similar models from other labs

Comparable on output type, weights, release date, price and context window. Prices are the lowest listed, USD per million tokens.

ModelLabReleasedContextInput $/1MOutput $/1MWeights
GLM-5.3Zhipu AI (Z.ai)14 Aug 20261M$0.07$1.40Open
LongCat-2.0Meituan30 Jun 20261M$0.30$1.20Open
GLM 5.3 FlashXZhipu AI (Z.ai)18 Sept 20261M$0.37$1.25–
MiniMax-M3MiniMax1 Jun 20261M$0.28$1.10Open
MiMo-V2.6-ProXiaomi (MiMo)22 Sept 20261M$0.43$0.87Open
DeepSeek V4 ProDeepSeek13 Aug 20261M$0.11$0.87Open

Change history

  • Input modalities changed from text, image to text, image, pdf
  • Max output changed from 524K to 1M
  • Context window changed from 524K to 1M
  • Input modalities changed from text, image, audio to text, image
  • Structured output changed from no to yes
  • Max output changed from 262K to 524K
  • Structured output changed from yes to no
  • Max output changed from 524K to 262K
  • Input modalities changed from text, image to text, image, audio
  • Context window changed from 1M to 524K
  • Input modalities changed from text, image, pdf to text, image
  • Max output changed from 1M to 524K
  • Input modalities changed from text, image, audio to text, image, pdf
  • Max output changed from 1M to 1M
  • Context window changed from 524K to 1M
  • Context window changed from 1M to 524K
  • Max output changed from 1M to 1M
  • Input modalities changed from text, image, pdf to text, image, audio
  • Input modalities changed from text, image, audio to text, image, pdf
  • Context window changed from 524K to 1M
  • Max output changed from 1M to 1M
  • Structured output changed from no to yes
  • Input modalities changed from text, image to text, image, audio
  • Max output changed from 1M to 1M
  • Input modalities changed from text, image, pdf to text, image

Sources