Noometry

Thinking Machines Lab, open weights

Inkling-Small

Inkling-Small by Thinking Machines Lab ranks 63rd of 354 ranked models on the Noometry Index as of October 2026, with a score of 46.5. Its strongest category is multimodal, where it ranks 62nd. API pricing starts at $0.45 per million input tokens and $1.20 per million output tokens, with a 524K-token context window.

Last verified

Noometry rank
#63 of 354
Index score
46.5
Evidence
Confirmed 33 results
Provider
Thinking Machines Lab
Released
July 15, 2026
Weights
Open weights
Reasoning
Yes
Context window
524K
Max output
1.05M
Input price
$0.45 / M
Output price
$1.20 / M
Blended price
$0.64 / M
Output speed
Not measured
Value
#81 of 219
Knowledge cutoff
Unknown
Input
text, image, audio
Hugging Face
thinkingmachines/Inkling-Small

Category scores

Each category score combines every public result we have in that category.

  1. Coding 43.6
  2. Reasoning 38.6
  3. Math 45.1
  4. Knowledge 48.2
  5. Multimodal 39.1
  6. Multilingual 51.7
  7. Instruction Following 73.8
  8. Long Context 42.7
  9. Writing & Preference 59.6
CategoryScoreRankResults
Coding43.6#853
Reasoning38.6#636
Math45.1#775
Knowledge48.2#773
Multimodal39.1#621
Multilingual51.7#1041
Instruction Following73.8#1141
Long Context42.7#1181
Writing & Preference59.6#1074

Strengths and weaknesses

Categories where Inkling-Small places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Reasoning38.6+15.0#63 of 350, top 18%
Math45.1+8.5#77 of 327, top 24%
Knowledge48.2+10.9#77 of 314, top 25%

Weakest categories

CategoryScorevs medianRank
Multimodal39.1+0.5#62 of 128, top 49%
Long Context42.7+1.8#118 of 296, top 40%
Instruction Following73.8+2.6#114 of 305, top 38%

Closest competitors

The models ranked just above and below Inkling-Small. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
GLM-5.1#5947.8$2.15—Compare
Kimi K2.6#6047.7$1.71—Compare
o3#6147.5$3.503Compare
Qwen3.6 Plus#6247.5$1.13—Compare
GPT-5 Pro#6446.4$41.255Compare
Grok 4.20 Multi-Agent#6546.2$1.56—Compare
GLM-5#6646.1$1.5523Compare
Qwen3.5 397B-A17B#6746.0$1.359Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
LMArena WebDev1409#68 of 113, top 61%Inkling SmallLMArena2026-10-08
SciCode48.7%#48 of 121, top 40%Epoch AI
LMArena Coding1451#86 of 294, top 30%Inkling SmallLMArena2026-10-08

Reasoning

BenchmarkScorePositionSettingSourceDate
ARC-AGI-233.1%highEpoch AI
ARC-AGI-24.9%lowEpoch AI
ARC-AGI-213.6%mediumEpoch AI
ARC-AGI-24%minimalEpoch AI
ARC-AGI-20%noneEpoch AI
ARC-AGI-240.1%#36 of 83, top 44%xhighEpoch AI
ARC-AGI-178%highEpoch AI
ARC-AGI-152.5%lowEpoch AI
ARC-AGI-167%mediumEpoch AI
ARC-AGI-148.5%minimalEpoch AI
ARC-AGI-15.5%noneEpoch AI
ARC-AGI-184%#37 of 83, top 45%xhighEpoch AI
CritPt8.3%#52 of 134, top 39%Epoch AI
Chess Puzzles18%#64 of 129, top 50%xhighEpoch AI2026-08-14
LMArena Hard Prompts1423#100 of 297, top 34%Inkling SmallLMArena2026-10-08
Mystery Game Puzzles6%#70 of 74, top 95%xhighEpoch AI2026-08-15
Epoch Capabilities Index150.15#51 of 213, top 24%Epoch AI2026-07-15

Math

BenchmarkScorePositionSettingSourceDate
FrontierMath (Tiers 1-3)46.3%#49 of 81, top 61%xhighEpoch AI2026-08-14
FrontierMath Tier 417.1%#48 of 63, top 77%xhighEpoch AI2026-08-14
OTIS Mock AIME 2024-202590%#52 of 173, top 31%xhighEpoch AI2026-08-14
ProofBench6%#68 of 77, top 89%Epoch AI
LMArena Math1459#53 of 285, top 19%Inkling SmallLMArena2026-10-08

Knowledge

BenchmarkScorePositionSettingSourceDate
GPQA Diamond88.5%#46 of 186, top 25%xhighEpoch AI2026-08-14
SimpleQA Verified19.1%#67 of 77, top 88%xhighEpoch AI2026-08-27
LMArena Expert1442#79 of 273, top 29%Inkling SmallLMArena2026-10-08

Multimodal

BenchmarkScorePositionSettingSourceDate
LMArena Vision1235#65 of 122, top 54%Inkling SmallLMArena2026-10-09

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1402#104 of 297, top 36%Inkling SmallLMArena2026-10-08
LMArena Chinese1465#82 of 285, top 29%Inkling SmallLMArena2026-10-08
LMArena French1436#85 of 223, top 39%Inkling SmallLMArena2026-10-08
LMArena German1405#90 of 231, top 39%Inkling SmallLMArena2026-10-08
LMArena Japanese1405#56 of 211, top 27%Inkling SmallLMArena2026-10-08
LMArena Korean1363#91 of 213, top 43%Inkling SmallLMArena2026-10-08
LMArena Russian1391#116 of 283, top 41%Inkling SmallLMArena2026-10-08
LMArena Spanish1428#85 of 226, top 38%Inkling SmallLMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
LMArena Instruction Following1399#107 of 298, top 36%Inkling SmallLMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
LMArena Longer Query1401#116 of 291, top 40%Inkling SmallLMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1414#106 of 297, top 36%Inkling SmallLMArena2026-10-08
LMArena Creative Writing1331#146 of 295, top 50%Inkling SmallLMArena2026-10-08
EQ-Bench Creative Writing1491#57 of 115, top 50%EQ-Bench
LMArena Multi-Turn1418#96 of 295, top 33%Inkling SmallLMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
deepinfra$0.45$1.20$0.102026-10-10
openrouter$0.45$1.20$0.102026-10-10

All Thinking Machines Lab API prices → Estimate your cost →

Compare Inkling-Small

Other Thinking Machines Lab models

Frequently asked questions

How good is Inkling-Small?

Inkling-Small by Thinking Machines Lab ranks 63rd of 354 ranked models on the Noometry Index as of October 2026, with a score of 46.5. Its strongest category is multimodal, where it ranks 62nd. API pricing starts at $0.45 per million input tokens and $1.20 per million output tokens, with a 524K-token context window.

How much does Inkling-Small cost?

Inkling-Small costs $0.45 per million input tokens and $1.20 per million output tokens on deepinfra, with cached input at $0.10.

What is Inkling-Small's context window?

Inkling-Small accepts up to 524K tokens of input and can write up to 1.05M tokens in one response.

Is Inkling-Small open source?

Yes. Inkling-Small's weights are downloadable from Hugging Face (thinkingmachines/Inkling-Small); check the license for commercial terms.

What are Inkling-Small's strengths and weaknesses?

Relative to other ranked models, Inkling-Small places best in reasoning, math, knowledge and lowest in multimodal, long context, instruction following.

What is Inkling-Small best at?

Its best category is multimodal, where it ranks 62nd on Noometry.