Ctrl K
Back to models

Thinking Machines

Inkling Small

  • Released
  • Open weights
  • Apache 2.0
Compare
Intelligence Index

Reasoning setting

25.7#41 of 83measured models
Measured · Artificial Analysis v4.3.2
API price from
$0.50 in / $1.20 out / 1M tokOpenRouter · BaseTen · FP8
Context window
1M tok

Performance profile

ReasoningRank #23 of 107
89.5%
GPQA Diamond
CodingRank #33 of 94
49.7%
SciCode
Cost efficiencyRank #15 of 67
286.8pts / $
Intelligence ÷ eval cost

Capabilities

Context window

Modalities

Measured setting

Integration

Reasoning trade-offs

Source (opens in a new tab)
Intelligence Index v4.3.2
Inkling Small
Cost per task (USD) · log scale
Inkling SmallReasoning
Intelligence Index
25.7
Cost / task
$0.0895
Evaluation time
2.2m

Pricing & access

API route
OpenRouter · BaseTen · FP8API
Tier & contextStandard

USD per 1M tokens

Input
$0.50
Output
$1.20
Cached input
$0.10

Run locally

GGUF
Context
Estimated memory

~167.8 GiB

RAM + VRAM

Estimate assumptions

One text chat at the selected context, with a 16-bit cache and app allowance. Operating system and image processing are extra; optimized caches can use less.

192 GiB RAMwithout GPU offload

Memory budgetOne text chat
Weights
151.4 GiB
Context + state
1.3 GiB16-bit cache
App allowance
15.1 GiB
Download
162.5 GB5 files · keep together+0.1 GB for images (opens in a new tab)
Disk space
~184 GBIncludes image file
Available software
Ollama (opens in a new tab)LM Studio (opens in a new tab)Jan (opens in a new tab)KoboldCpp (opens in a new tab)llama.cpp (opens in a new tab)
Get GGUF (opens in a new tab)
Set up in Ollama · Q4_K_M · 8K context

Save the following as Modelfile. The context matches the estimate above.

FROM hf.co/unsloth/Inkling-Small-GGUF:Q4_K_M
PARAMETER num_ctx 8192

Then run these commands in your terminal:

ollama create llmcurrent-local -f Modelfile
ollama run llmcurrent-local

Ollama import instructions (opens in a new tab) · For LM Studio or Jan, open the app’s model search, enter unsloth/Inkling-Small-GGUF, choose Q4_K_M, and set context to 8,192 tokens in the app’s model settings. Software links above open each app’s setup guide.

Images need extra memoryvLLM setup (opens in a new tab)

Benchmarks

Full leaderboardSource (opens in a new tab)

Humanity’s Last Exam

Frontier reasoning

Inkling Small · Reasoning#46 of 126

  1. 1Claude Opus 5.5Max61.4%
  2. 2Claude Fable 5.1Max59.1%
  3. 3Gemini 4 ArgonHigh57.1%
  5. 43LongCat 2.0Reasoning33.7%
  6. 44Agnes 2.5 Pro AlphaReasoning33.6%
  7. 45Hy3Reasoning33.5%
  8. 46Inkling SmallReasoning33.3%
  9. 47K2 Horizon 375B A23BReasoning32.0%
  10. 48InklingExtra high31.9%
  11. 49Nemotron 3.5 Super VLReasoning30.7%

Peers use their best measured result on this benchmark.

Alternatives

Each model at its highest-scoring measured setting · Intelligence Index v4.3.2 and its evaluation cost per task.