Inkling
Reasoning effort
As of 29 Oct 2026, Inkling by Thinking Machines scores 49.8 on the BenchLeader Index (581 ranked configurations), with its best results in long context (63) and weakest in coding (35). It costs $2.57 per 1M tokens blended ($1.00 in / $4.05 out), and streams about 210 tokens/s (Artificial Analysis). Released 15 Jul 2026; 14 independent results so far.
Key facts (as of 29 Oct 2026)
- Inkling costs $2.57 per 1M tokens blended ($1.00 in / $4.05 out) (list price, 29 Oct 2026).
- Inkling streams about 210 tokens/s (Artificial Analysis, 29 Oct 2026).
- Inkling was released on 15 Jul 2026 (Thinking Machines).
In detail. Inkling is a Thinking Machines open-weights reasoning model, released 15 Jul 2026. Its best configuration (xhigh reasoning effort) ranks #335 of 581 on the BenchLeader Index at 49.8 ±5.9, in the lower half. The ± is the point: 285 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 long context (63) and lowest in coding (35). At $1.76 per million tokens blended it is pricier than most ranked models. Output speed of 206 tokens per second puts it in the fastest quarter, with a first answer in 11.5 s. It has been measured at 5 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 15 Jul 2026.
- Blended price
- $1.76/M
- $1.00 in · $4.05 out
- Output speed
- 206 tok/s
- measured by Artificial Analysis
- First answer
- 11 s
- first token 0.32 s
- Context
- 66k
- Full answer
- 14 s
- median, reasoning included
- Cost per run
- $1.35
- one full Intelligence Index run
- Released
- 15 Jul 2026
- Overall index50
- Reasoning47
- Coding35
- Agents & tools56
- Maths45
- Knowledge46
- Multimodal56
- Long context63
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| high | – | – | 79 tok/s | 0.32 s | $0.0022 | Coding 38 |
| xhighbest | 49.8 | #335 | 206 tok/s | 11 s | $0.0016 | Agents & tools 56 · Coding 35 · Knowledge 46 · Long context 63 · Maths 45 · Multimodal 56 · Reasoning 47 |
| max | – | – | 79 tok/s | 0.32 s | $0.0022 | |
| thinking | – | – | 79 tok/s | 0.32 s | $0.0022 | |
| not stated | 47.8 | #383 | 79 tok/s | 0.32 s | $0.0022 | Agents & tools 31 · Coding 46 · Composite 28 · Human preference 59 · Knowledge 60 · Maths 54 · Reasoning 53 |
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.
Metadata
| Benchmark | xhigh | Source | Trend |
|---|---|---|---|
| AA Openness Indexnot in index | 38.9#34 | Artificial Analysis |
Reasoning
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| GPQA Diamond | 88.3%#64 | – | Epoch AI Benchmarking Hub | |
| SimpleBencheffort not stated | – | 50.0%#45 | SimpleBench | |
| Arena Hard Promptseffort not stated | – | 1462#91 | Arena (LMArena) | |
| LiveBench 2.0 Reasoningnot in indexeffort not stated | – | 78.3%#25 | LiveBench | |
| LiveBench 1.0 Reasoningnot in index | 78.3%#56 | – | LiveBench | |
| GPQA Diamond (AA)not in index | 87.2%#103 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 31.9%#150 | – | Artificial Analysis | |
| HLE-Diamond (AA) | 10.6%#93 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in indexeffort not stated | – | 87.1%#39 | Vals AI | |
| ARC-AGI-1effort not stated | – | 79.5%#122 | ARC Prize | |
| ARC-AGI-2effort not stated | – | 36.5%#123 | ARC Prize | |
| CritPtnot in index | 5.4%#161 | – | Artificial Analysis | |
| CritPt v2 hard subset (AA) | 18.0%#73 | – | Artificial Analysis | |
| Chess Puzzlesnot in index | 21.0%#88 | – | Epoch AI Benchmarking Hub | |
| LMCAnot in index | 37.6%#117 | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | 87.5%#100 | – | Epoch AI Benchmarking Hub |
Coding
| Benchmark | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|
| SciCode | – | 47.0%#138 | – | SciCode | |
| WeirdML | 32.3%#132 | – | – | WeirdML | |
| FrontierCodeeffort not stated | – | – | 14.0%#42 | Cognition | |
| Arena Codingeffort not stated | – | – | 1494#92 | Arena (LMArena) | |
| Arena WebDeveffort not stated | – | – | 1415#90 | Arena (LMArena) | |
| LiveBench 2.0 Codingnot in indexeffort not stated | – | – | 64.0%#26 | LiveBench | |
| LiveBench 1.0 Codingnot in index | – | 71.0%#60 | – | LiveBench | |
| SciCode (AA)not in index | – | 47.0%#144 | – | Artificial Analysis | |
| LiveCodeBencheffort not stated | – | – | 85.5%#31 | Vals AI | |
| IOIeffort not stated | – | – | 14.9%#44 | Vals AI | |
| SWE-bench (Vals)not in indexeffort not stated | – | – | 77.6%#34 | Vals AI | |
| Code Migrationnot in indexeffort not stated | – | – | 11.8%#62 | Vals AI | |
| Code Migration v1.1not in indexeffort not stated | – | – | 8.8%#28 | Vals AI | |
| ProgramBenchnot in indexeffort not stated | – | – | 0.0%#32 | Vals AI | |
| Vibe Code Bench 1-100not in indexeffort not stated | – | – | 7.3%#25 | Vals AI | |
| Vibe Code Bench v1.1not in indexeffort not stated | – | – | 19.2%#91 | Vals AI | |
| ALE-Benchnot in indexeffort not stated | – | – | 946#64 | Epoch AI Benchmarking Hub | |
| FrontierSWEnot in index | – | 4.1%#20 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | – | 1.0%#137 | – | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | – | 55.1%#110 | – | Artificial Analysis | |
| FrontierCode 1.1 (AA)not in index | – | 15.8%#70 | – | Artificial Analysis | |
| FrontierCode 2.0 (AA) | – | 3.9%#74 | – | Artificial Analysis |
Agents & tools
| Benchmark | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|
| APEX-Agentseffort not stated | – | – | 33.8%#52 | Mercor | |
| Arena Agenteffort not stated | – | – | -15.8#48 | Arena (LMArena) | |
| LiveBench Long Agenticnot in indexeffort not stated | – | – | 8.5%#29 | LiveBench | |
| LiveBench Browsernot in indexeffort not stated | – | – | 12.0%#29 | LiveBench | |
| LiveBench Agentic Codingnot in index | 49.4%#45 | – | – | LiveBench | |
| Terminal-Bench 2.1 (Vals)effort not stated | – | – | 47.6%#63 | Vals AI | |
| MCP Atlas | 76.0%#17 | – | – | Scale AI SEAL | |
| τ³-bench Bankingnot in index | – | 25.0%#18 | – | τ²-bench | |
| GDPval-AA v2.1not in index | 1081#89 | – | – | Artificial Analysis | |
| τ³-Banking (AA) | 29.1%#71 | – | – | Artificial Analysis | |
| Analyst Agent (AA)not in index | 23.8%#35 | – | – | Artificial Analysis | |
| CyberBenchnot in indexeffort not stated | – | – | 66.8%#30 | Vals AI | |
| Finance Agent v2not in indexeffort not stated | – | – | 46.6%#55 | Vals AI | |
| Finance Agent v2.1not in indexeffort not stated | – | – | 49.9%#33 | Vals AI | |
| Harvey's Legal Agent Benchmark v1.1not in indexeffort not stated | – | – | 3.8%#29 | Vals AI | |
| Harvey's Legal Agent Benchmark v1not in indexeffort not stated | – | – | 2.1%#48 | Vals AI | |
| Legal Research Benchnot in indexeffort not stated | – | – | 28.4%#58 | Vals AI | |
| SkillsBenchnot in indexeffort not stated | – | – | 26.1%#33 | Vals AI | |
| SREBenchnot in indexeffort not stated | – | – | 0.0%#32 | Vals AI | |
| Tax Agent Benchnot in indexeffort not stated | – | – | 15.3%#57 | Vals AI | |
| Terminal-Bench 4.0 (Vals)not in indexeffort not stated | – | – | 0.5%#54 | Vals AI | |
| AutomationBenchnot in index | 5.0%#175 | – | – | Artificial Analysis | |
| GDP.pdf (AA)not in index | 12.8%#118 | – | – | Artificial Analysis | |
| Harvey LAB-AA v1.1not in index | 0.0%#23 | – | – | Harvey | |
| EnterpriseOps-Gymnot in index | 38.0%#33 | – | – | Artificial Analysis | |
| AA-Briefcase v1.1not in index | 833#117 | – | – | Artificial Analysis | |
| Terminal-Bench-Science 0.1 (AA)not in index | 0.0%#90 | – | – | Artificial Analysis |
Maths
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| FrontierMath Tiers 1–3 | 33.3%#72 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | 4.9%#67 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 88.9%#77 | – | Epoch AI Benchmarking Hub | |
| ProofBencheffort not stated | – | 0.0%#80 | Vals AI | |
| LiveBench Mathematicsnot in index | 88.4%#43 | – | LiveBench | |
| Arena Mathseffort not stated | – | 1480#40 | Arena (LMArena) |
Knowledge
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| SimpleQA Verified | 40.3%#50 | – | Epoch AI Benchmarking Hub | |
| LiveBench 2.0 Data Analysisnot in indexeffort not stated | – | 72.8%#20 | LiveBench | |
| LiveBench Researchnot in indexeffort not stated | – | 59.9%#27 | LiveBench | |
| LiveBench 1.0 Data Analysisnot in index | 72.8%#44 | – | LiveBench | |
| AA-Omniscience v1.1 | -1.5#62 | – | Artificial Analysis | |
| MMLU-Proeffort not stated | – | 86.3%#42 | Vals AI | |
| LegalBencheffort not stated | – | 83.0%#63 | Vals AI | |
| CorpFineffort not stated | – | 68.6%#7 | Vals AI | |
| TaxEvaleffort not stated | – | 75.3%#19 | Vals AI | |
| Excel Modeling Benchmarknot in indexeffort not stated | – | 38.4%#60 | Vals AI | |
| Excel Modeling Benchmark v1.1not in indexeffort not stated | – | 41.0%#35 | Vals AI | |
| MedCodenot in indexeffort not stated | – | 41.2%#56 | Vals AI | |
| MedScribenot in indexeffort not stated | – | 85.4%#28 | Vals AI | |
| Public Benefits Bench v1.1not in indexeffort not stated | – | 58.5%#36 | Vals AI | |
| AA-Omniscience v1.1: accuracynot in index | 39.3%#54 | – | Artificial Analysis | |
| AA-Omniscience v1.1: non-hallucinationnot in index | 32.8%#72 | – | Artificial Analysis |
Instruction following
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| LiveBench Languagenot in index | 73.5%#56 | – | LiveBench | |
| LiveBench Instruction Followingnot in index | 70.1%#30 | – | LiveBench | |
| Arena Instruction Followingnot in indexeffort not stated | – | 1426#104 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1442#91 | Arena (LMArena) | |
| EQ-Bench 4 | 1212#16 | EQ-Bench | |
| Arena Creative Writingnot in index | 1386#131 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1444#99 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1441#108 | Arena (LMArena) |
Multimodal
| Benchmark | xhigh | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| LiveBench Designnot in indexeffort not stated | – | – | 58.3%#30 | LiveBench | |
| MMMU-Pro | 73.5%#129 | – | – | Artificial Analysis | |
| MMMU-Pro (Vals)not in indexeffort not stated | – | – | 78.5%#50 | Vals AI | |
| MortgageTaxnot in indexeffort not stated | – | – | 64.0%#48 | Vals AI | |
| AudioMCnot in index | – | 56.6%#4 | – | Scale AI SEAL | |
| AudioMC (text output)not in index | – | 56.6%#4 | – | Scale AI SEAL | |
| Code2Video-AAnot in index | 467#26 | – | – | Artificial Analysis | |
| AA-3Dnot in index | 583#22 | – | – | Artificial Analysis |
Long context
| Benchmark | xhigh | Source | Trend |
|---|---|---|---|
| AA-LCR v1.1 | 77.3%#139 | Artificial Analysis | |
| MLCRnot in index | 6.1%#142 | Artificial Analysis |
Composite
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | 148.4#69 | Epoch AI Benchmarking Hub | |
| LiveBench 2.0effort not stated | – | 47.0%#28 | LiveBench | |
| LiveBench 1.0not in index | 71.9%#52 | – | LiveBench | |
| AA Intelligence Index v5.0not in index | 12.5#73 | – | Artificial Analysis | |
| AA Intelligence Index v4.3.2not in index | 25.0#174 | – | Artificial Analysis | |
| AA Progress Index v1not in index | 281.4#169 | – | Artificial Analysis | |
| Vals Indexnot in indexeffort not stated | – | 29.0#29 | Vals AI | |
| Vals Finance Indexnot in indexeffort not stated | – | 38.3%#33 | Vals AI | |
| Vals Multimodal Indexnot in indexeffort not stated | – | 49.4%#26 | Vals AI | |
| Vals Router Indexnot in indexeffort not stated | – | 29.0%#21 | Vals AI |
Safety & honesty
| Benchmark | xhigh | Source | Trend |
|---|---|---|---|
| FORTRESSnot in index | 25.0%#34 | Scale AI SEAL |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Together | 123 tok/s | 0.18 s | $1.00 | $4.05 | 524k | – |
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 at $0.374 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0016 | $0.0014 | 12.9 s |
| Summarize a 30-page report | 12,000 / 600 | $0.014 | $0.0088 | 14.4 s |
| Code edit | 6,000 / 1,500 | $0.012 | $0.0093 | 18.8 s |
| Agentic coding session | 60,000 / 4,000 | $0.076 | $0.048 | 30.9 s |
| Structured extraction | 2,000 / 200 | $0.0028 | $0.0019 | 12.5 s |
More from Thinking Machines
See also
Questions people ask
- What are Inkling's benchmark scores?
- As of 29 Oct 2026, Inkling (xhigh effort) scores 49.8 ±5.9 on the BenchLeader Index, strongest in long context (63) and weakest in coding (35). That rests on 14 independent benchmark results, and every score on this page links to its publisher.
- How does Inkling rank?
- Inkling (xhigh effort) is #335 of 581 ranked configurations on 29 Oct 2026. 285 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 69.1.
- How much does Inkling cost per 1M tokens?
- As of 29 Oct 2026, Inkling costs $1.00 per 1M input tokens and $4.05 per 1M output tokens, $1.76 blended (three input tokens per output token).
- When was Inkling released?
- Thinking Machines released Inkling on 15 Jul 2026. BenchLeader has results for it at 5 reasoning-effort settings.
Data as of 29 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Inkling: benchmarks, pricing, speed and rank”, data as of 29 Oct 2026, https://www.benchleader.com/models/inkling.