AI Tools Review

Inkling 2 Small

By Thinking Machines

Released: 2026-11-05

LLM
Thinking Machines
Open Weights
Free
New

Inkling 2 Small is Thinking Machines's open-weight AI model, released 05/11/2026. Inkling 2 Small is a 594B-parameter mixture-of-experts model with 17B active, accepting image input and supporting variable effort. Thinking Machines said the full weights would be released under Apache 2.0 after external safety testing and red-teaming (the page notes weights on Hugging Face on 12 November 2026). Its comparison table covers GLM-5.3-Flash, MiMo V2.6, MiniMax-M3.1-Flash, DeepSeek-V4.1-Flash-1017, Muse Spark Open, GLM 5.3, Mistral Large 4, Kimi K3, GPT-6 Luna, Gemini 3.8 Flash and Claude Haiku 5.5. The page prints some figures inconsistently (Intelligence Index 31 against 30.6, DesignArena 1331 against 1340). Our article on this model is the sibling Inkling 2 Mini piece.

Visit Inkling 2 Small

AI-Powered

Leverages advanced AI technology to deliver cutting-edge capabilities and results.

Fast & Efficient

Optimized performance ensures quick results without compromising on quality.

Purpose-Built

Specifically designed for llms tasks and workflows.

How does Inkling 2 Small score on independent benchmarks?

Inkling 2 Small ranks #20 of 28 models on the Artificial Analysis Intelligence Index (v5.0.1) with a score of 30.6, as measured on 22 November 2026.

Intelligence Index
30.6
AA v5.0.1
Coding Agent Index
n/a
Cost per task
$1.18
AA Intelligence Index task
Price per 1M tokens
$0.3 / $1
input / output

Where does Inkling 2 Small rank against other models?

How does Inkling 2 Small do on individual tests?

Evaluation (run by Artificial Analysis)Inkling 2 SmallBest of any model
Terminal-Bench 4.013.1%67.2%
FrontierCode 2.025.4%48.8%
Humanity's Last Exam44.1%74.3%
GDPval-AA (Elo)16301866

How has Inkling 2 Small's score changed over time?

We have tracked Inkling 2 Small on Intelligence Index v5.0.1 since 22/11/2026. A point is added each day the score, coding index or price changes, so this chart fills in over time.

What does the Inkling 2 Small system card report?

Inkling 2 Small is a 594B-parameter mixture-of-experts model with 17B active, accepting image input and supporting variable effort. Thinking Machines said the full weights would be released under Apache 2.0 after external safety testing and red-teaming (the page notes weights on Hugging Face on 12 November 2026). Its comparison table covers GLM-5.3-Flash, MiMo V2.6, MiniMax-M3.1-Flash, DeepSeek-V4.1-Flash-1017, Muse Spark Open, GLM 5.3, Mistral Large 4, Kimi K3, GPT-6 Luna, Gemini 3.8 Flash and Claude Haiku 5.5. The page prints some figures inconsistently (Intelligence Index 31 against 30.6, DesignArena 1331 against 1340). Our article on this model is the sibling Inkling 2 Mini piece.

These are Thinking Machines's own published results (self-reported, with the vendor's harness and settings), so they are not directly comparable with other labs' numbers. Source: Inkling 2 Small (05/11/2026). Last checked 22/11/2026.

Coding

BenchmarkInkling 2 SmallGLM 5.3Kimi K3GPT-6 Luna
FrontierCode v2.0Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory25.4%n/an/an/a
DeepSWE v1.1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory72.1%69%68.5%66.6%
SWE-Bench Pro v2 (Hard)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory86.3%84.3%88.2%88.7%
SWE-Bench Pro v2 (Full)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory97.7%n/an/an/a
SWE-Atlas v1.1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory61.9%n/an/an/a
Vibe Code Bench v1.1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory81.6%n/an/an/a
Terminal Bench Science v0.1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory5.7%9.5%8.1%8.6%
PostTrainBench v1.2Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory31.9%n/an/an/a
PMPP-Hard (AUC, 0 to 10h)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory494.27 AUCn/an/an/a

Agents & tool use

BenchmarkInkling 2 SmallGLM 5.3Kimi K3GPT-6 Luna
AA-Briefcase v1.1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory1527 Elon/an/an/a
GDPval-AA v2.1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory1630 Elo165715431444
JobBenchEffort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory70.5%n/an/an/a
EnterpriseOps-GymEffort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory46.8%n/an/an/a
Finance Agent v2.1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory67.5%n/an/an/a
Apex-Accounting v1.1 (pass@1)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory14.5%n/an/an/a
BrowseComp (with compaction)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory90.2%75.1%88.4%83.6%
DeepSearchQA (with compaction)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory90.9%n/an/an/a

Knowledge & science

BenchmarkInkling 2 SmallGLM 5.3Kimi K3GPT-6 Luna
HLE-DiamondEffort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory28.7%17.9%26.9%22.2%
CritPt v2 (Full)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory55.7%n/an/an/a
CritPt v2 (Hard)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory35%n/an/an/a
AA-Omniscience scoreEffort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory16.1 pointsn/an/an/a
AA-Omniscience non-hallucination rateEffort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory73.3%n/an/an/a

Reasoning & maths

BenchmarkInkling 2 SmallGLM 5.3Kimi K3GPT-6 Luna
ARC-AGI-1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory96.5%n/an/an/a
ARC-AGI-2Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory81.7%n/a60.4%59.3%

Long context

BenchmarkInkling 2 SmallGLM 5.3Kimi K3GPT-6 Luna
AA-LCR v1.1Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory83.3%79.7%88.7%83.3%
LongBench v2Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory70.6%n/an/an/a

Multimodal

BenchmarkInkling 2 SmallGLM 5.3Kimi K3GPT-6 Luna
MMMU Pro (Standard 10)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory81.6%n/a80.5%79.7%
PerceptionBenchEffort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory64.7%n/an/an/a
ScreenSpot Pro (original)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory83.8%n/a73.8%81.1%
ScreenSpot Pro (with Python)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory92%n/a86.1%87.7%
CharXiv RQ (original)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory85.6%n/an/an/a
CharXiv RQ (with Python)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory89.1%n/an/an/a
BabyVision (original)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory68.6%n/an/an/a
BabyVision (with Python)Effort 0.99 (xhigh), temperature 1.0, 1M-token maximum trajectory93.6%n/an/an/a

Other

BenchmarkInkling 2 SmallGLM 5.3Kimi K3GPT-6 Luna
DesignArena WebsiteTable value; the chart on the same page shows 13401331 Elon/an/an/a
Artificial Analysis Intelligence IndexBenchmark table value; the page intro rounds to 3130.6 pointsn/an/an/a

Safety & honesty

BenchmarkInkling 2 SmallGLM 5.3Kimi K3GPT-6 Luna
StrongREJECTHigher is better97.1 pointsn/an/an/a
FORTRESS average risk scoreLower is better16.2 pointsn/an/an/a
FORTRESS overrefusal scoreLower is better20.4 pointsn/an/an/a

What did safety testing find for Inkling 2 Small?

Safety level: Thinking Machines' own evaluations; no external framework level stated

  • DistressBench 82.0 (higher is better) and PatchEval 79.6 (higher is better); CyberBench v1.1 72.7.
  • Internal misbehaviour rate 14.9, covert misbehaviour 11.8 and destructive actions 4.7 (all lower is better).
  • Apollo Research covert-action-with-goal result is 0.
  • The page states Inkling 2 Small trails GPT-6.1 Sol on its alignment evaluations.

Independent figures: Artificial Analysis, refreshed daily on our AI model benchmarks page. Full analysis: our Inkling 2 Small deep dive.

Thinking Machines Model Timeline

Inkling 2 SmallCurrent

Not published as a product limit; evaluations used a 1M-token maximum trajectory (256K for RL training) context

Specifications

pricing$0.3/$1 per 1M tokens (input/output)
context WindowNot published as a product limit; evaluations used a 1M-token maximum trajectory (256K for RL training)