Data as of November 21, 2026 · How the score is built
Inkling 2 Mini
Decision readingInkling 2 Mini scores 55.9 out of 100 and ranks #70 of 215. This profile shows 45 source-displayable benchmark rows; its strongest eligible category is Multilingual at #39. API pricing is $0.15 input and $0.5 output per million tokens, with cached input at $0.015.
Email me when its price, rank or availability changes.
Capability
55.9/100
field median 48.1#70 of 215 ranked models
Public
#70of 215
Verified —
Price
$0.15input / $0.50 output
input median $0.60cached $0.015 · blended $0.33
Speed
Not measured
field median 93.5 tok/sTime to first token not measured
Context
1Mtokens
field median 256,000Maximum output length is tracked separately
Decision snapshot
Where this model sits in the ranked field, its strongest published evidence, and what to validate before choosing it. The marker compares it with the current ranked catalog, not with an absolute quality threshold.
Strongest published evidence
Multilingual ranks #39. A well-rounded choice across a range of tasks.
Validate before choosing
45 published rows leave some tracked benchmark slots empty. Independent runtime speed has not been measured.
Source-linked · 45 displayable benchmark rows
Is Inkling 2 Mini right for your job?
Its place on the LLM Selector’s job rankings, the same order the job pages publish. Describe your own job to see where it stands for that.
Where Inkling 2 Mini ranks by benchmark score:
We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy
Category score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #61 of 179Percentile 66thWeight 22%8 benchmarksVerified | 59.1 | 8 benchmarks | Verified | |||
| CodingRank #67 of 179Percentile 63rdWeight 20%5 benchmarksVerified | 53.9 | 5 benchmarks | Verified | |||
| ReasoningRank #58 of 161Percentile 64thWeight 17%2 benchmarksVerified | 57.8 | 2 benchmarks | Verified | |||
| MultimodalRank #62 of 119Percentile 48thWeight 12%13 benchmarksVerified | 55.5 | 13 benchmarks | Verified | |||
| KnowledgeRank #79 of 158Percentile 50thWeight 12%3 benchmarksVerified | 51.9 | 3 benchmarks | Verified | |||
| MultilingualRank #39 of 87Percentile 56thWeight 7%Not listedUnlisted sources | 54.7 | Not listed | Unlisted sources | |||
| Inst. FollowingWeight 5%1 benchmarkVerified | Score pending | 1 benchmark | Verified | |||
| MathWeight 5%1 benchmarkVerified | Score pending | 1 benchmark | Verified |
45 of 937 tracked benchmark slots have displayable evidence · bars run 0–100
Coverage detailsHow much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic8/8 verified
- Coding5/5 verified
- Reasoning2/2 verified
- Multimodal13/13 verified
- Knowledge3/3 verified
- MultilingualNot measured
- Inst. Following1/1 verified
- Math1/1 verified
Capability shape
Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.
Inkling 2 Mini category percentile values
- Agentic66th percentile
- Coding63rd percentile
- Reasoning64th percentile
- Multimodal48th percentile
- Knowledge50th percentile
- Multilingual56th percentile
- Instruction followingNot eligible
- MathNot eligible
The dashed outline is median of 6 nearest peers.
Eligible category ranks
- Agentic#61/179
- Coding#67/179
- Reasoning#58/161
- Multimodal#62/119
- Knowledge#79/158
- Multilingual#39/87
- Inst. FollowingNot ranked
- MathNot ranked
Benchmark ledger
Every result says who ran the test. For this model, Thinking Machines Lab ran all 73. Coding opens by default. The tick marks the best source-verified result in the catalog; provisional leaders do not set the reference.
Coding5 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| DeepSWE | 48.2% | Best verified: Muse Wonder Full · 79.0% | Gap30.8 behind | Reported by Thinking Machines Lab | Counts5% ref. weight |
| SWE Multilingual | 73.7% | Best verified: Claude Fable 5.2 · 96.6% | Gap22.9 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| FrontierCode 2.0 | 15.1% | Best verified: Claude Fable 5.2 · 50.2% | Gap35.1 behind | Reported by Thinking Machines Lab | Shown only |
| Terminal-Bench 2.1 | 71.6% | Best verified: SWE-2 · 92.8% | Gap21.2 behind | Reported by Thinking Machines Lab | Shown only |
| SWE-bench VerifiedSoftware Engineering Benchmark Verified | 79.2% | Best verified: Claude Opus 5 · 96% | Gap16.8 behind | Reported by Thinking Machines Lab | Shown only |
Agentic8 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| Terminal-Bench 4.0 | 5.60% | Best verified: Claude Sonnet 5.5 · 70.60% | Gap65 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| Terminal-Bench-Science 0.1 | 1.4% | Best verified: GPT-6.2 Astra · 77.1% | Gap75.7 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| BrowseComp | 79% | Best verified: Atria Dawn Preview · 92.5% | Gap13.5 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| MCP Atlas | 82.7% | Best verified: Muse Spark Open · 89.2% | Gap6.5 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| JobBench | 49.6% | Best verified: Muse Wonder Full · 75.3% | Gap25.7 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| DeepSearchQA | 85.3% | Best verified: Atria Dawn Preview · 96.0% | Gap10.7 behind | Reported by Thinking Machines Lab | Counts2% ref. weight |
| Terminal-Bench 2.1 | 71.6% | Best verified: SWE-2 · 92.8% | Gap21.2 behind | Reported by Thinking Machines Lab | Shown only |
| EnterpriseOps-GymEnterpriseOps-Gym (provider runs) | 42.00% | Best verified: Inkling 2 Mini · 42.00% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
Reasoning2 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| LongBench v2 | 66.9% | Best verified: Inkling 2 Small · 70.6% | Gap3.7 behind | Reported by Thinking Machines Lab | Counts6% ref. weight |
| MRCR v2 256K-512KOpenAI MRCR v2 8-needle 256K-512K | 46.4% | Best verified: GPT-6 Astra · 100.0% | Gap53.6 behind | Reported by Thinking Machines Lab | Shown only |
Multimodal13 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| ScreenSpot Pro | 76.5% | Best verified: GPT-6 Astra · 92.7% | Gap16.2 behind | Reported by Thinking Machines Lab | Counts5% ref. weight |
| MMMU-ProMassive Multi-discipline Multimodal Understanding Pro | 74.9% | Best verified: GPT-6.1 Terra · 85.5% | Gap10.6 behind | Reported by Thinking Machines Lab | Counts4% ref. weight |
| MathVision | 89.0% | Best verified: Qwen3.8 Max · 95.2% | Gap6.2 behind | Reported by Thinking Machines Lab | Counts4% ref. weight |
| CharXivCharXiv Reasoning | 88.3% | Best verified: Qwen3.8 Max · 93.5% | Gap5.2 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| CharXiv w/o toolsCharXiv Reasoning without tools | 80.6% | Best verified: Kimi K3.1 Pro · 89% | Gap8.4 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| BabyVision | 53.4% | Best verified: Qwen3.8 Max · 82.0% | Gap28.6 behind | Reported by Thinking Machines Lab | Counts2% ref. weight |
| ZeroBench | 17.0% | Best verified: Muse Spark Open · 50.0% | Gap33 behind | Reported by Thinking Machines Lab | Counts2% ref. weight |
| PerceptionBenchPerceptionBench (Internal) | 57.5% | Best verified: Kimi K3.1 Pro · 68.3% | Gap10.8 behind | Reported by Thinking Machines Lab | Shown only |
| OmniDocBench 1.6OmniDocBench v1.6 | 92.8% | Best verified: Inkling 2 Mini · 92.8% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
| BabyVision w/ PythonBabyVision with Python | 88.8% | Best verified: Kimi K3.1 Pro · 95.4% | Gap6.6 behind | Reported by Thinking Machines Lab | Shown only |
| Chartography (no tools)Chartography without tools | 22.1% | Best verified: Claude Fable 5.2 · 73.4% | Gap51.3 behind | Reported by Thinking Machines Lab | Shown only |
| Chartography (tools)Chartography with image and code tools | 72.5% | Best verified: Claude Fable 5.2 · 91.6% | Gap19.1 behind | Reported by Thinking Machines Lab | Shown only |
| ZeroBench w/ PythonZeroBench_main with Python | 34.0% | Best verified: Kimi K3.1 Pro · 63.0% | Gap29 behind | Reported by Thinking Machines Lab | Shown only |
Knowledge3 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| HLE w/o toolsHumanity's Last Exam without tools | 30.7% | Best verified: Claude Fable 5.2 · 73.2% | Gap42.5 behind | Reported by Thinking Machines Lab | Counts4% ref. weight |
| GPQA-DGPQA Diamond | 86.6% | Best verified: GPT-6.2 Astra · 97.0% | Gap10.4 behind | Reported by Thinking Machines Lab | Counts3% ref. weight |
| GPQAGraduate-Level Google-Proof Q&A | 86.6% | Best verified: GPT-6.2 Astra · 97% | Gap10.4 behind | Reported by Thinking Machines Lab | Shown only |
Inst. Following1 row
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| MultiChallenge (provider reports)MultiChallenge: provider-reported runs | 50.00% | Best verified: Mercury 3 · 60.35% | Gap10.4 behind | Reported by Thinking Machines Lab | Shown only |
Math1 row
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| Apex | 34.4% | Best verified: Hy4 preview · 74.2% | Gap39.8 behind | Reported by Thinking Machines Lab | Counts15% ref. weight |
External signals12 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| PMPP-HardPMPP-Hard GPU kernel optimization (AUC, 0-10 hours) | 455.82 task-hours | Best verified: Inkling 2 Small · 494.27 task-hours | Gap38.4 behind | Reported by Thinking Machines Lab | Shown only |
| ScreenSpot Pro w/ PythonScreenSpot Pro with Python | 87.7% | Best verified: Inkling 2 Small · 91.5% | Gap3.8 behind | Reported by Thinking Machines Lab | Shown only |
| ComplexConstraintsComplexConstraints (Surge) | 33.4% | Best verified: Inkling 2 Small · 44.5% | Gap11.1 behind | Reported by Thinking Machines Lab | Shown only |
| Diligence BenchDiligenceBench (Thoughtful), best harness | 48.6% | Best verified: Inkling 2 Small · 58.1% | Gap9.5 behind | Reported by Thinking Machines Lab | Shown only |
| StrongREJECTStrongREJECT (no jailbreak) | 97.8% | Best verified: Inkling 2 Small · 97.8% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
| BVBBVB (Vision Interpretation and Creation) | 68.4% | Best verified: Inkling 2 Mini · 68.4% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
| Intelligence per WattIntelligence per Watt (local AI energy efficiency) | 0.703 /J | Best verified: Inkling 2 Mini · 0.703 /J | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
| CritPt v2 (Full)CritPt v2 Full set (research-level physics) | 31.4% | Best verified: Inkling 2 Mini · 31.4% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
| CritPt v2 (Hard)CritPt v2 Hard subset (research-level physics) | 17.0% | Best verified: Inkling 2 Mini · 17.0% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
| GraphWalks BFS 256KGraphWalks BFS at 256K context | 69.7% | Best verified: Inkling 2 Mini · 69.7% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
| FORTRESS (adversarial, Public)FORTRESS Public adversarial safety split | 89.5% | Best verified: Inkling 2 Mini · 89.5% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
| FORTRESS (benign, Public)FORTRESS Public benign over-refusal split | 96.4% | Best verified: Inkling 2 Mini · 96.4% | GapHighest listed | Reported by Thinking Machines Lab | Shown only |
Other lab-reported results28 rowsAnother metric, split or setting, or a test without a page
| Benchmark | Score | On a 0–100 scale | Who ran it | Use in ranking |
|---|---|---|---|---|
| SAGE v1.1Score · Reported in the Inkling 2 Mini model card. | 85.9% | Reported by Thinking Machines Lab | Shown only | |
| ARC-AGI-1Accuracy · Reported in the Inkling 2 Mini launch table and model card. | 84.5% | Reported by Thinking Machines Lab | Shown only | |
| AA-LCR v1.1Score · Reported in the Inkling 2 Mini model card. | 80.0% | Reported by Thinking Machines Lab | Shown only | |
| SWE-bench Pro v2 (Hard)Tasks resolved · 51-task Hard split using mini-swe-agent v2.4.6. | 77.9% | Reported by Thinking Machines Lab | Shown only | |
| Vibe Code Bench v1.1Score · openhands (vals); reported in the Inkling 2 Mini launch table. | 77.9% | Reported by Thinking Machines Lab | Shown only | |
| MuirBenchOverall accuracy · Reported in the Inkling 2 Mini launch table; not part of the original MuirBench ICLR 2025 paper table. | 77.5% | Reported by Thinking Machines Lab | Shown only | |
| Excel Modeling Benchmark v1.1Score · Reported in the Inkling 2 Mini model card. | 66.0% | Reported by Thinking Machines Lab | Shown only | |
| AA-Omniscience v1.1 (non-hallucination)Non-hallucination rate · Reported in the Inkling 2 Mini launch table. | 63.4% | Reported by Thinking Machines Lab | Shown only | |
| Finance Agent v2.1Score · Reported in the Inkling 2 Mini model card. | 56.5% | Reported by Thinking Machines Lab | Shown only | |
| GroundingMEOverall accuracy · Reported in the Inkling 2 Mini model card; not yet evaluated on the official GroundingME CVPR 2026 leaderboard. | 55.5% | Reported by Thinking Machines Lab | Shown only | |
| Vals Finance IndexIndex score · Reported in the Inkling 2 Mini launch table. | 50.1% | Reported by Thinking Machines Lab | Shown only | |
| ARC-AGI-2Accuracy · Reported in the Inkling 2 Mini launch table and model card. | 49.3% | Reported by Thinking Machines Lab | Shown only | |
| Vals Index (v2.2)Overall score · Reported in the Inkling 2 Mini launch table under Third Party Index. | 46.5% | Reported by Thinking Machines Lab | Shown only | |
| AA-AnalystAgentScore · Reported in the Inkling 2 Mini launch table. | 46.3% | Reported by Thinking Machines Lab | Shown only | |
| AutomationBench-AA v1.0.6Task completion · Reported in the Inkling 2 Mini launch table. | 41.8% | Reported by Thinking Machines Lab | Shown only | |
| SWE-bench Multimodal v2Tasks resolved · mini-swe-agent (vals); reported in the Inkling 2 Mini model card. | 37.9% | Reported by Thinking Machines Lab | Shown only | |
| Code Migration v1.1Score · pi (vals); reported in the Inkling 2 Mini model card. | 23.2% | Reported by Thinking Machines Lab | Shown only | |
| MysteryMechanismScore · Reported in the Inkling 2 Mini model card. | 23.0% | Reported by Thinking Machines Lab | Shown only | |
| ProofBench v1.1Score · Reported in the Inkling 2 Mini model card. | 18.0% | Reported by Thinking Machines Lab | Shown only | |
| GDP.pdfScore · Reported in the Inkling 2 Mini launch table under Vision Interpretation and Creation. | 16.8% | Reported by Thinking Machines Lab | Shown only | |
| HLE-DiamondAccuracy · Reported in the Inkling 2 Mini launch table. | 15.2% | Reported by Thinking Machines Lab | Shown only | |
| Tax Agent BenchScore · Reported in the Inkling 2 Mini model card. | 14.1% | Reported by Thinking Machines Lab | Shown only | |
| HoneyBench v0.2Reward-hacking rate (lower is better) · Reported in the Inkling 2 Mini model card before inclusion on Goodhart Labs' HoneyBench v0.2 leaderboard. | 56.7% | Reported by Thinking Machines Lab | Shown only | |
| GDPVal-AA v2.1Elo rating · Reported in the Inkling 2 Mini launch table. | 1319 | Reported by Thinking Machines Lab | Shown only | |
| AA-Briefcase v1.1Elo rating · Reported in the Inkling 2 Mini launch table. | 1263 | Reported by Thinking Machines Lab | Shown only | |
| Code2Video-AARating · Reported in the Inkling 2 Mini model card. | 870 | Reported by Thinking Machines Lab | Shown only | |
| AA Index (v5.0)Intelligence Index · Reported in the Inkling 2 Mini launch table under Third Party Index. | 20.1 | Reported by Thinking Machines Lab | Shown only | |
| AA-Omniscience v1.1Omniscience score · Reported in the Inkling 2 Mini launch table. | -8.6 | Reported by Thinking Machines Lab | Shown only |
Outlined bars are lab-reported. Filled bars are benchmark-run or independent.
What it costs to get this score
Published API price against the public score. The x-axis uses a log scale; the dashed path marks models that are not beaten by a cheaper, higher-scoring option. Price uses average of published input and output rates.
Inkling 2 Mini · 55.9 score · $0.33 blended per million tokens
The chart opens on the current model. Scroll horizontally to inspect the full price axis.
Horizontal: blended price per million tokens, log scale · Vertical: public score
Lineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- thinkingmachines/Inkling-2-MiniTinker models and pricing
- Context window
- 1MTinker models and pricing
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- text, imageInkling 2 Mini model card
- Output modalities
- textInkling 2 Mini model card
- Parameters
- Not sourced yet
- Availability
- Open weights released under Apache 2.0 on Hugging Face (thinkingmachines/Inkling-2-Mini) as a ~20GB native NVFP4/MXFP8 checkpoint, supported locally on Apple Silicon (32GB+ RAM via Ollama or Exo Labs), NVIDIA DGX Spark, SGLang, vLLM, Unsloth, llama.cpp, NVIDIA Dynamo, and TokenSpeed. Hosted inference and fine-tuning are available on Tinker (64K and 256K training tiers, 1M sampling context) and the Tinker Playground. Accepts text and image inputs and outputs text.
- Cloud regions
- Not tracked yet
- Lifecycle
- Current
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Published at $0.015 per million cached input tokensTinker models and pricing
- Self-host
- Open weights available; hardware estimate not sourced
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
Inkling 2 Mini ranks #70 of 215 on the public leaderboard with a score of 55.9/100. It does not yet have enough sourced coverage for a verified position.
Inkling 2 Mini is a open weight model with a 1M context window. No explicit reasoning mode is documented in this profile.
Open weights released under Apache 2.0 on Hugging Face (thinkingmachines/Inkling-2-Mini) as a ~20GB native NVFP4/MXFP8 checkpoint, supported locally on Apple Silicon (32GB+ RAM via Ollama or Exo Labs), NVIDIA DGX Spark, SGLang, vLLM, Unsloth, llama.cpp, NVIDIA Dynamo, and TokenSpeed. Hosted inference and fine-tuning are available on Tinker (64K and 256K training tiers, 1M sampling context) and the Tinker Playground. Accepts text and image inputs and outputs text.
Official exact-value snapshot from Thinking Machines Lab’s November 20, 2026 Inkling 2 Mini launch post and model card. Inkling 2 Mini is a 24-layer sparse MoE looped transformer (26.4B total parameters, 3.4B unique active parameters, 5.4B effective active parameters per token across three passes of the middle 12 layers, 8 of 128 experts routed per layer) trained and deployed natively in NVFP4 with the last four layers in MXFP8 (~20GB checkpoint), under Apache 2.0. Exact rows with a compatible BenchLM key are stored below; Terminal-Bench 2.1 (pi v0.83.0 harness) stays display-only out of cohort calibration. Third-party board scores, split-specific rows, and benchmarks without a catalog key are recorded in the lab-reported ledger.
Inkling 2 Mini sits in the Inkling family with Inkling, Inkling 2 Small, Inkling-Small. 45 of 937 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Multilingual at #39, while its lowest eligible position is Knowledge at #79. a well-rounded choice across a range of tasks.
Last updated November 21, 2026. Runtime fields remain blank until a sourced snapshot exists.
Deployment options
Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.
Published weights are available, but BenchLM does not yet have a sourced parameter and VRAM profile for this exact model. Hardware cost estimates stay unavailable until that sizing record is complete.
Estimate VRAM from known parametersQuestions
How does Inkling 2 Mini perform overall in AI benchmarks?
Inkling 2 Mini ranks #70 out of 215 models on the public BenchAlign leaderboard, with a score of 55.9/100. Its evidence status is Estimated, and this profile shows 45 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
Is Inkling 2 Mini good for knowledge and understanding?
Inkling 2 Mini ranks #79 out of 158 eligible models for knowledge and understanding, with a public category score of 51.9/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Inkling 2 Mini good for coding and programming?
Inkling 2 Mini ranks #67 out of 179 eligible models for coding and programming, with a public category score of 53.9/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Inkling 2 Mini good for mathematics?
Inkling 2 Mini has source-displayable benchmark coverage for mathematics, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is Inkling 2 Mini good for reasoning and logic?
Inkling 2 Mini ranks #58 out of 161 eligible models for reasoning and logic, with a public category score of 57.8/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Inkling 2 Mini good for agentic tool use and computer tasks?
Inkling 2 Mini ranks #61 out of 179 eligible models for agentic tool use and computer tasks, with a public category score of 59.1/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Inkling 2 Mini good for multimodal and grounded tasks?
Inkling 2 Mini ranks #62 out of 119 eligible models for multimodal and grounded tasks, with a public category score of 55.5/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Inkling 2 Mini good for instruction following?
Inkling 2 Mini has source-displayable benchmark coverage for instruction following, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is Inkling 2 Mini open source?
Inkling 2 Mini is an open-weight model from Thinking Machines Lab. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.
Which sibling models are related to Inkling 2 Mini?
Inkling 2 Mini belongs to the Inkling family. Related tracked variants include Inkling, Inkling 2 Small, Inkling-Small. A sibling link indicates shared lineage or a documented configuration relationship; it does not mean the variants have identical pricing, context limits, benchmark evidence, or deployment behavior. Compare before switching.
Does Inkling 2 Mini have full benchmark coverage on BenchLM?
No. Inkling 2 Mini currently has 50 source-displayable rows. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of Inkling 2 Mini?
Inkling 2 Mini has a documented context window of 1M. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.