Inkling 2 Small
1 cheaper ranked model scores higher: GPT-5.6 Luna (58.5, $0.450).
As of 6 Nov 2026, Inkling 2 Small by Thinking Machines scores 57.4 on the BenchLeader Index (587 ranked configurations), with its best results in agents & tools (69) and weakest in reasoning (53). API pricing: $0.300 per 1M input tokens and $1.00 per 1M output tokens ($0.475 blended at 3:1), list price on 6 Nov 2026. Released 5 Nov 2026.
Key facts (as of 6 Nov 2026)
- Inkling 2 Small API pricing: $0.300 per 1M input tokens and $1.00 per 1M output tokens ($0.475 blended at 3:1) (list price, 6 Nov 2026).
- Inkling 2 Small was released on 5 Nov 2026 (Thinking Machines).
In detail. Inkling 2 Small is a Thinking Machines proprietary reasoning model, released 5 Nov 2026. Its best configuration ranks #146 of 587 on the BenchLeader Index at 57.4 ±4.9, in the upper half. The ± is the point: 260 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in agents & tools (69) and lowest in reasoning (53). At $0.475 per million tokens blended it is cheaper than most ranked models. Last measured 5 Nov 2026.
- Blended price
- $0.475/M
- $0.300 in · $1.00 out
- via Artificial Analysis
- Output speed
- –
- First answer
- –
- Context
- 1.0M
- via Artificial Analysis
- Full answer
- –
- Cost per task
- $1.18
- 21 min a task, AA Index
- Tokens per task
- 150k
- 59% of it reasoning
- Released
- 5 Nov 2026
- Overall index57
- Reasoning53
- Coding56
- Agents & tools69
- Knowledge55
- Multimodal64
- Long context67
Versions
Thinking Machines has shipped 2 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| NewInkling 2 Smallthis page | 5 Nov 2026 | 57.4 | #146 |
| Inkling Small | 15 Jul 2026 | 52.1 | #283 |
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | not stated | Source |
|---|---|---|
| Humanity's Last Exam (AA)not in index | 44.1%#54 | Artificial Analysis |
| HLE-Diamond (AA) | 28.7%#47 | Artificial Analysis |
| CritPt v2 hard subset (AA) | 35.0%#55 | Artificial Analysis |
Coding
| Benchmark | not stated | Source |
|---|---|---|
| SciCode (AA)not in index | 52.3%#90 | Artificial Analysis |
| ProgramBenchnot in index | 0.5%#18 | Vals AI |
| SWE-bench Multimodalnot in index | 43.1%#12 | Vals AI |
| Vibe Code Bench v1.1not in index | 81.5%#28 | Vals AI |
| Terminal-Bench 4.0 (AA)not in index | 13.1%#92 | Artificial Analysis |
| Terminal-Bench 2.1 (AA)not in index | 81.3%#41 | Artificial Analysis |
| FrontierCode 2.0 (AA) | 25.4%#42 | Artificial Analysis |
Agents & tools
| Benchmark | not stated | Source |
|---|---|---|
| GDPval-AA v2.1not in index | 1630#24 | Artificial Analysis |
| τ³-Banking (AA) | 40.4%#29 | Artificial Analysis |
| Analyst Agent (AA)not in index | 51.3%#9 | Artificial Analysis |
| CyberBenchnot in index | 72.7%#16 | Vals AI |
| Finance Agent v2.1not in index | 67.5%#4 | Vals AI |
| Harvey's Legal Agent Benchmark v1.1not in index | 21.7%#5 | Vals AI |
| Legal Research Benchnot in index | 46.6%#17 | Vals AI |
| SREBenchnot in index | 1.9%#26 | Vals AI |
| Tax Agent Benchnot in index | 38.9%#12 | Vals AI |
| Terminal-Bench 4.0 (Vals)not in index | 12.6%#41 | Vals AI |
| Terminal-Bench Sciencenot in index | 2.9%#33 | Vals AI |
| AutomationBenchnot in index | 45.6%#107 | Artificial Analysis |
| GDP.pdf (AA)not in index | 25.2%#39 | Artificial Analysis |
| EnterpriseOps-Gymnot in index | 46.8%#12 | Artificial Analysis |
| AA-Briefcase v1.1not in index | 1527#31 | Artificial Analysis |
| Terminal-Bench-Science 0.1 (AA)not in index | 5.7%#66 | Artificial Analysis |
Knowledge
| Benchmark | not stated | Source |
|---|---|---|
| AA-Omniscience v1.1 | 16.1#51 | Artificial Analysis |
| BioMysteryBenchnot in index | 70.0%#14 | Vals AI |
| Excel Modeling Benchmark v1.1not in index | 69.5%#18 | Vals AI |
| AA-Omniscience v1.1: accuracynot in index | 33.8%#76 | Artificial Analysis |
| AA-Omniscience v1.1: non-hallucinationnot in index | 73.3%#21 | Artificial Analysis |
Multimodal
| Benchmark | not stated | Source |
|---|---|---|
| MMMU-Pro | 81.6%#46 | Artificial Analysis |
| SAGEnot in index | 88.6%#8 | Vals AI |
| Code2Video-AAnot in index | 1020#16 | Artificial Analysis |
| AA-3Dnot in index | 1021#10 | Artificial Analysis |
Long context
| Benchmark | not stated | Source |
|---|---|---|
| AA-LCR v1.1 | 83.3%#21 | Artificial Analysis |
Composite
| Benchmark | not stated | Source |
|---|---|---|
| AA Intelligence Index v5.0not in index | 30.6#52 | Artificial Analysis |
| Vals Indexnot in index | 56.1#19 | Vals AI |
| Vals Finance Indexnot in index | 61.5%#7 | Vals AI |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0004 | – | – |
| Summarize a 30-page report | 12,000 / 600 | $0.0042 | – | – |
| Code edit | 6,000 / 1,500 | $0.0033 | – | – |
| Agentic coding session | 60,000 / 4,000 | $0.022 | – | – |
| Structured extraction | 2,000 / 200 | $0.0008 | – | – |
More from Thinking Machines
Compare Inkling 2 Small with…
See also
Embed or cite
Show where Inkling 2 Small ranks on a model card or README. The badge redraws itself with every daily update, carries the data date and links back to this page. To see ranks like this on OpenRouter and Hugging Face model pages, get the Chrome extension.
[](https://www.benchleader.com/models/inkling-2-small)Cite this
The ranking changes daily, so a reference carries the data date. APA style:
BenchLeader. (2026, November 6). Inkling 2 Small on the BenchLeader Index (method v1.4) [Data set]. Retrieved November 6, 2026, from https://www.benchleader.com/models/inkling-2-small
BibTeX:
@misc{benchleader_inkling_2_small_20261106,
author = {{BenchLeader}},
title = {{Inkling 2 Small on the BenchLeader Index (method v1.4)}},
year = {2026},
month = nov,
howpublished = {\url{https://www.benchleader.com/models/inkling-2-small}},
note = {Data as of 2026-11-06}
}Every badge style, the embeddable top-10 widget and how to cite the Index
Questions people ask
- What are Inkling 2 Small's benchmark scores?
- As of 6 Nov 2026, Inkling 2 Small scores 57.4 ±4.9 on the BenchLeader Index, strongest in agents & tools (69) and weakest in reasoning (53). That rests on independent benchmark results only, and every score on this page links to its publisher.
- How does Inkling 2 Small rank?
- Inkling 2 Small is #146 of 587 ranked configurations on 6 Nov 2026. 260 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 69.3.
- How much does Inkling 2 Small cost per 1M tokens?
- As of 6 Nov 2026, Inkling 2 Small costs $0.300 per 1M input tokens and $1.00 per 1M output tokens, $0.475 blended (three input tokens per output token).
- Is Inkling 2 Small better than DeepSeek V4 Pro?
- As of 6 Nov 2026, DeepSeek V4 Pro (high) and Inkling 2 Small are neck and neck on quality (57.5 vs 57.4 on the BenchLeader Index, inside the ±6.6 margin of error), and Inkling 2 Small costs 2.1× less, which makes it the better buy for most workloads. The head-to-head page has the benchmark-by-benchmark detail.
- When was Inkling 2 Small released?
- Thinking Machines released Inkling 2 Small on 5 Nov 2026.
Data as of 6 Nov 2026. Compare with another model.
Cite as: BenchLeader, “Inkling 2 Small: benchmarks, pricing, speed and rank”, data as of 6 Nov 2026, https://www.benchleader.com/models/inkling-2-small.