Thinking Machines: Models Intelligence, Performance & Price

Name: Intelligence
Creator: Artificial Analysis
License: https://artificialanalysis.ai/docs/legal/Terms-of-Use.pdf

Thinking Machines

This analysis is intended to support you in choosing the best model provided by Thinking Machines for your use-case.

Most Intelligent

Inkling 2 Small

Inkling Small

Inkling (xhigh)

Intelligence Index

Total 3 models

Lowest Time per Task

Inkling Small

5.3 min

Inkling (xhigh)

10.7 min

Inkling 2 Small

28.5 min

Time per Intelligence Index task (minutes)

Total 3 models

Lowest Price

Inkling 2 Small

$0.18

Inkling Small

$0.22

Inkling (xhigh)

$0.72

Blended price (per 1M tokens)

Total 3 models

Thinking Machines offers 3 models, each with different intelligence, performance, and pricing characteristics. Below is a comparison of the key metrics across models.

For intelligence, the top models on Thinking Machines are Inkling 2 Small (31), Inkling Small (13), and Inkling (xhigh) (12).
For time per task, Inkling Small (5.3 min), Inkling (xhigh) (10.7 min), and Inkling 2 Small (28.5 min) finish an Intelligence Index task fastest on Thinking Machines.
For latency, Inkling Small (15.14s), Inkling (xhigh) (21.00s), and Inkling 2 Small (68.69s) offer the lowest time to first answer token.
For pricing, Inkling 2 Small ($0.18), Inkling Small ($0.22), and Inkling (xhigh) ($0.72) offer the lowest blended prices per 1M tokens.
For context window size, Inkling 2 Small (262k), Inkling (xhigh) (262k), and Inkling Small (256k) support the largest context windows on Thinking Machines.
Inkling 2 Small provides the best balance of intelligence and cost-effectiveness. For the lowest time per task, Inkling Small is the top choice.

Intelligence

Artificial Analysis Intelligence Index · Higher is better

Time per Task

Inference time (minutes) per Intelligence Index task · Lower is better

Price

USD per 1M tokens (blended) · Lower is better

Intelligence

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index v5.0.1 incorporates 9 evaluations: AA-Briefcase v1.1, AutomationBench-AA, Terminal-Bench 4.0, FrontierCode-AA 2.0, AA-Omniscience v1.1, GDP.pdf, HLE-Diamond, CritPt v2 (Hard Subset), and Terminal-Bench-Science 0.1

Artificial Analysis Intelligence Index v5.0.1 includes: AA-Briefcase v1.1, AutomationBench-AA, Terminal-Bench 4.0, FrontierCode-AA 2.0, AA-Omniscience v1.1, GDP.pdf, HLE-Diamond, CritPt v2 (Hard Subset), Terminal-Bench-Science 0.1. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Evaluations

Evaluations measured independently by Artificial Analysis · AA-PersonalityBench scores are percentile ranks from 1 to 100

AA-Briefcase v1.1

Agentic knowledge work, Elo

AA-Omniscience v1.1 Accuracy

Knowledge

AA-Omniscience v1.1 Non-Hallucination Rate

1 - hallucination rate

AutomationBench-AA

Agentic SaaS workflows

Terminal-Bench-Science 0.1

Agentic scientific research workflows in a terminal

FrontierCode-AA 2.0

Agentic software engineering in open-source repositories

HLE-Diamond

Reasoning & knowledge

GDP.pdf

Professional document reasoning, All-pass

Terminal-Bench 4.0

Agentic coding & terminal use

CritPt v2 (Hard Subset)

Physics reasoning

AA-AnalystAgent

Quantitative analysis on spreadsheets & documents

Harvey LAB-AA v1.1

Legal agentic work, Hallucination-Gated All-Pass Rate

EnterpriseOps-Gym-AA

Agentic business operations

GDPval-AA v2.1

Agentic real-world work tasks, Elo

AA-3D

Agentic interactive 3D experiences, Elo

Code2Video-AA

Product and brand videos designed in code, Elo

ITBench-AA

Kubernetes incident root-cause analysis

No data available

MLCR-AA

Medical long context reasoning

MMMU-Pro

Visual reasoning

AA-PersonalityBench: Expressiveness Score

Rich, varied, well-crafted language

AA-PersonalityBench: Organization Score

Well-structured, scannable, answer-first

AA-PersonalityBench: Transparency Score

Truthful, acknowledges uncertainty, non-sycophantic

AA-PersonalityBench: Slop Score

Use of filler and generic language

While model intelligence generally translates across use cases, specific evaluations may be more relevant for certain use cases.

Intelligence Index vs. Price

Blended at 7:2:1 (cache-input-output) · USD per 1M tokens (blended)

Most attractive quadrant

While higher intelligence models are typically more expensive, they do not all follow the same price-quality curve.

Pricing

Cost per Intelligence Index Task

Weighted average cost (USD) per Intelligence Index task · Segmented by token type · Lower is better

Weighted average cost per Intelligence Index task. Each evaluation’s cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

Intelligence Index vs. Price

Blended at 7:2:1 (cache-input-output) · USD per 1M tokens (blended)

Most attractive quadrant

While higher intelligence models are typically more expensive, they do not all follow the same price-quality curve.

Speed

Average of the single-request and multi-turn agentic workloads measured

Time per Intelligence Index Task

Output tokens divided by end-to-end output speed (minutes per Intelligence Index task) · Lower is better

The weighted average time (minutes) per Artificial Analysis Intelligence Index task: output tokens per task divided by output speed, weighted by each benchmark's weight in the Intelligence Index.

Speed and latency metrics average each endpoint's runs over the past 5 days, and the error rate pools them. We run the benchmark every 6 hours, and endpoints of deprecated models every 18 hours.

Quality

Context Window

Context window: tokens limit · Higher is better

Larger context windows are relevant to RAG (Retrieval Augmented Generation) LLM workflows which typically involve reasoning and information retrieval of large amounts of data.

Maximum number of combined input & output tokens. Output tokens commonly have a significantly lower limit (varied by model).

Latency

Average of the single-request and multi-turn agentic workloads measured

Time to First Answer Token

Seconds until the first answer token after any reasoning · Average of the single-request and multi-turn agentic workloads measured · Lower is better

Time to first answer token received, in seconds, after API request sent. For reasoning models, this includes the 'thinking' time of the model before providing an answer. For models which do not support streaming, this represents time to receive the completion.

Speed and latency metrics average each endpoint's runs over the past 5 days, and the error rate pools them. We run the benchmark every 6 hours, and endpoints of deprecated models every 18 hours.

Reliability

Share of requests the endpoint failed, across all the single-request and multi-turn agentic workloads measured

Endpoint Error Rate

Share of requests the endpoint failed with an error, a rate limit or a timeout · Across all the single-request and multi-turn agentic workloads measured · Lower is better

Share of requests over the past 5 days that the endpoint failed with an error, a rate limit or a timeout. Failures on our side are left out of both counts. Across several workloads, the failures and requests are pooled before dividing.

Speed and latency metrics average each endpoint's runs over the past 5 days, and the error rate pools them. We run the benchmark every 6 hours, and endpoints of deprecated models every 18 hours.

Caching

Cache Hit Rate

Median across multi-turn agentic session runs in the last four weeks, updated Oct 26, 2026 · Higher is better

The share of cacheable input tokens served from the provider's prompt cache, measured during multi-turn agentic sessions. We regularly measure each endpoint and report a representative rate based on recent measurements. Higher values mean more of the input is billed at the cheaper cache-hit price, where the provider offers one, which lowers the effective cost of agentic workloads. For further details, see our methodology page.

Cost per Task vs. Cache Hit Rate

Weighted average cost (USD) per Intelligence Index task · Cache hit rate: median of the last four weeks, updated Oct 26, 2026

Most attractive quadrant

Pareto line


								Further Analysis
Inkling 2 Small	31	$1.18	30.9	66.64	0.0%	262k	Open	Model Provider
Inkling Small	13	$0.20	5.5	19.04	0.0%	256k	Open	Model Provider
Inkling (xhigh)	12	$1.35	10.6	21.20	5.5%	262k	Open	Model Provider

Key definitions

Average cost per task in the index. Costs are split by input, cache hit, cache write, reasoning, and answer token pricing where canonical token counts are available.

Frequently Asked Questions

Common questions about Thinking Machines

Thinking Machines offers 3 models that we track: Inkling 2 Small, Inkling Small, and Inkling (xhigh).

The most intelligent model available on Thinking Machines is Inkling 2 Small with an Intelligence Index score of 31.

The model with the lowest time per task on Thinking Machines is Inkling Small at 5.3 minutes per Intelligence Index task. Time per task is the output tokens per task divided by end-to-end output speed, so lower means a task finishes sooner.

The model with the lowest time to first answer token on Thinking Machines is Inkling Small at 15.14s. Lower latency means faster initial response time.

The most affordable model on Thinking Machines by blended price is Inkling 2 Small at $0.18 per 1M tokens (7:2:1 cache hit/input/output ratio).

Prices on Thinking Machines vary up to 4.0x across models, from $0.18 per 1M tokens for Inkling 2 Small to $0.72 per 1M tokens for Inkling (xhigh).

Yes, Thinking Machines offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.

Yes, all 3 models on Thinking Machines support JSON mode for structured output.

Yes, all 3 models on Thinking Machines support function calling (tool use).

Yes, Thinking Machines offers 3 reasoning models: Inkling 2 Small, Inkling Small, and Inkling (xhigh). Reasoning models use extended thinking to work through complex problems before providing an answer.

Yes, all 3 models on Thinking Machines are open weight models.

Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.

When choosing a model on Thinking Machines, consider: intelligence (for quality-sensitive tasks), time per task (how quickly the model completes an Intelligence Index task), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and features like context window size, JSON mode, or function calling support.

Thinking Machines: Models Intelligence, Performance & Price

Most Intelligent

Lowest Time per Task

Lowest Price

Intelligence

Time per Task

Price

Intelligence

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index

Evaluations

Intelligence Evaluation Relevance

Artificial Analysis Intelligence Index

Intelligence Index vs. Price

Price-Quality Variance

Pricing

Cost per Intelligence Index Task

Cost per Intelligence Index Task

Intelligence Index vs. Price

Price-Quality Variance

Speed

Time per Intelligence Index Task

Time per Intelligence Index Task

Time Period

Quality

Context Window

Context Window for RAG

Context Window

Latency

Time to First Answer Token

Time to First Answer Token

Time Period

Reliability

Endpoint Error Rate

Endpoint Error Rate

Time Period

Caching

Cache Hit Rate

What Cache Hit Rate Shows

Cost per Task vs. Cache Hit Rate

Cost per Intelligence Index Task

What Cache Hit Rate Shows

Key definitions

Index Cost per Task

Frequently Asked Questions

What models does Thinking Machines offer?

What is the most intelligent model on Thinking Machines?

Which model has the lowest time per task on Thinking Machines?

Which model has the lowest latency on Thinking Machines?

What is the cheapest model on Thinking Machines?

How much do prices vary across models on Thinking Machines?

Is Thinking Machines OpenAI compatible?

Does Thinking Machines support JSON mode?

Does Thinking Machines support function calling?

Does Thinking Machines offer reasoning models?

Does Thinking Machines offer open weight models?

Does Thinking Machines's performance change over time?

How do I choose a model on Thinking Machines?