thinkingmachines · Open weights
Inkling Small
The direct-answer version of Thinking Machines' 276B-parameter open-weights multimodal Mixture-of-Experts model with 12B active parameters. It accepts text and images, and is designed for coding, tool use, instruction following, and general conversational work.
- Tools
- Vision
- Audio input
- Input
- Text, Images, Audio
- Output
- Text
- Discovery tier
- Open weights
Pricing
USD per 1M tokens- Input
- $0.45
- Output
- $1.20
- Cache read
- $0.10
Reference prices as published by the model provider. API example
Specifications
- Context window
- 524.3K
- Max output
- 32.8K
- Average output (7 days)
- 577 tokens
- Longer than 54% of 465 models
Side by side
Inkling Small compared
Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.
| Inkling SmallthinkingmachinesThis model | Grok 4.20xAIRemove | |
|---|---|---|
| Provider | thinkingmachines | xAI |
| Discovery tier | Open weights | Balanced |
| Added | Jul 31, 2026 | Mar 31, 2026 |
| Limits | ||
| Context window | 524.29K | 2MBest |
| Max output | 32.77K | 131.07KBest |
| Average output | 577 tokens | 679 tokens |
| Pricing, USD per 1M tokens | ||
| Input | $0.45Best | $1.25 |
| Output | $1.20Best | $2.50 |
| Cache read | $0.10Best | $0.20 |
| Performance | ||
| Throughput | — | 91.6 tok/s |
| Time to first token | — | 2.0 s |
| Request success (24h) | — | 99.8% |
| Benchmark indexes | ||
| Intelligence Index | 25.7 | 25.7 |
| Coding Index | 52.9 | — |
| Capabilities | ||
| Reasoning | No | Yes |
| Tools | Yes | Yes |
| Vision | Yes | Yes |
| Structured output | No | Yes |
| Audio input | Yes | No |
| Reasoning effort | — | — |
Public evaluations
Benchmarks
Independent evaluation results, with how each score ranks among the models in the Noumen catalog.
- Inkling Small
- Grok 4.20
Intelligence Index
- Inkling Small25.7
- Grok 4.2025.7
Inkling Small: Better than 62% of 326 scored models
Indexes
Coding Index
Composite of the coding evaluations
- Inkling Small: 52.9
- Grok 4.20: —
Agentic work
AutomationBench-AA
Workflow automation, penalized for guardrail violations
- Inkling Small: 4.8%
- Grok 4.20: —
AutomationBench-AA tasks completed
Workflows finished end to end without a guardrail violation
- Inkling Small: 0.6%
- Grok 4.20: —
Harvey LAB-AAv1.1
Legal tasks passed in full, with no material hallucination
- Inkling Small: 0.0%
- Grok 4.20: —
AA-Briefcasev1.1
Agentic knowledge work, rated head to head
- Inkling Small: 911 Elo
- Grok 4.20: —
GDPval-AAv2.1
Economically valuable occupational tasks, rated head to head
- Inkling Small: 1,124 Elo
- Grok 4.20: —
Document reasoning
GDP.pdf
Professional reasoning over PDF documents, all-pass rate
- Inkling Small: 11.2%
- Grok 4.20: —
AA-LCRv1.1
Reasoning across long documents
- Inkling Small: 75.7%Top
- Grok 4.20: 69.0%
MLCR-AA
Long-context reasoning over medical records
- Inkling Small: 7.8%
- Grok 4.20: —
Reasoning and knowledge
Humanity’s Last Exam
Expert-level questions across academic fields
- Inkling Small: 33.3%
- Grok 4.20: 34.5%Top
GPQA Diamond
Graduate-level science questions
- Inkling Small: 89.5%
- Grok 4.20: 91.1%Top
IFBench
Following precise output instructions
- Inkling Small: —
- Grok 4.20: 81.2%
Coding
Terminal-Benchv4.0
Practical software engineering in a terminal
- Inkling Small: 1.0%
- Grok 4.20: —
SciCode
Scientific computing in Python
- Inkling Small: 49.7%
- Grok 4.20: —
Legacy evaluations
Terminal-Bench Hard
Earlier terminal task suite
- Inkling Small: —
- Grok 4.20: 37.9%
τ²-Bench Telecom
Tool use in a customer-support simulation
- Inkling Small: —
- Grok 4.20: 93.0%
AA-LCR (unversioned)
Long-context reasoning, earlier grading
- Inkling Small: 75.7%Top
- Grok 4.20: 69.0%
Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis
- Inkling Small
- Grok 4.20
| Leaderboard | Inkling Small | Grok 4.20 |
|---|---|---|
| Vision Style Control · Entity Recognition | Not in top 20 | #7 · 1238.1 |
Ranks among Noumen catalog models on LMArena leaderboards. Only top-20 placements are listed.LMArena
- Inkling Small
- Grok 4.20
| Leaderboard | Inkling Small | Grok 4.20 |
|---|---|---|
| FrontierMath v2 (Tiers 1–3) | Not in top 20 | #14 · 44.9% |
| GPQA Diamond | Not in top 20 | #12 · 89.3% |
Ranks among Noumen catalog models on Epoch AI leaderboards. Only top-20 placements are listed.Epoch AI
- Inkling Small
- Grok 4.20
| Leaderboard | Inkling Small | Grok 4.20 |
|---|---|---|
| ARC-AGI-3 (Default harness) | Not in top 20 | #20 · 0.1% |
Ranks among Noumen catalog models on ARC Prize leaderboards. Only top-20 placements are listed.ARC Prize
Developers
Use it through the API
Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.
curl https://www.noumen.xyz/api/v1/chat/completions \
-H "Authorization: Bearer $NOUMEN_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "thinkingmachines/Inkling-Small",
"messages": [{ "role": "user", "content": "Explain this stack trace." }]
}'import os
from openai import OpenAI
client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
model="thinkingmachines/Inkling-Small",
messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)thinkingmachines
Related models
Other thinkingmachines models, starting with the same family.