Inkling-Small
Best configuration ranks #184 of 610 on the BenchLeader Index at 55.0 ±4.2. Last measured 8 Sept 2026. Released 15 Jul 2026.
- Blended price
- $0.638/M
- $0.450 in · $1.20 out
- Output speed
- 150 tok/s
- First answer
- 15 s
- first token 0.33 s
- Context
- 1M
- Overall index55
- Reasoning60
- Coding56
- Agents & tools49
- Maths34
- Knowledge60
- Human preference59
- Multimodal58
- Long context64
- Composite62
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. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | – | – | 150 tok/s | 15 s | $0.0006 | Reasoning 31 |
| minimal | – | – | 150 tok/s | 15 s | $0.0006 | Reasoning 39 |
| low | – | – | 150 tok/s | 15 s | $0.0006 | Reasoning 40 |
| medium | – | – | 150 tok/s | 15 s | $0.0006 | Reasoning 45 |
| high | – | – | 150 tok/s | 15 s | $0.0006 | Reasoning 52 |
| xhigh | 48.5 | #322 | 150 tok/s | 15 s | $0.0006 | Knowledge 32 · Maths 52 · Reasoning 59 |
| defaultbest | 55.0 | #184 | 150 tok/s | 15 s | $0.0005 | Agents & tools 49 · Coding 56 · Composite 62 · Human preference 59 · Knowledge 60 · Long context 64 · Maths 34 · Multimodal 58 · Reasoning 60 |
Benchmark results
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 | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | – | – | – | – | – | 88.5%#48 | – | Epoch AI Benchmarking Hub | |
| LMArena Hard Prompts | – | – | – | – | – | – | 1430#120 | LMArena | |
| GPQA Diamond (AA)not in index | – | – | – | – | – | – | 89.5%#74 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | – | – | – | – | – | – | 33.3%#92 | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | – | – | – | 83.6%#55 | Vals AI | |
| ARC-AGI-1 | 5.5%#174 | 48.5%#124 | 52.5%#121 | 67.0%#99 | 78.0%#85 | 84.0%#77 | – | ARC Prize | |
| ARC-AGI-2 | 0.0%#174 | 4.0%#127 | 4.9%#120 | 13.6%#99 | 33.1%#86 | 40.1%#79 | – | ARC Prize |
Coding
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SciCode | – | – | – | – | – | – | 48.7%#68 | SciCode | |
| LMArena Coding | – | – | – | – | – | – | 1477#94 | LMArena | |
| LMArena WebDev | – | – | – | – | – | – | 1406#67 | LMArena | |
| SciCode (AA)not in index | – | – | – | – | – | – | 49.6%#72 | Artificial Analysis | |
| LiveCodeBench | – | – | – | – | – | – | 85.9%#25 | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | – | – | – | 82.2%#20 | Vals AI |
Agents & tools
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Agent | – | – | – | – | – | – | -6.9#36 | LMArena | |
| Terminal-Bench 2.1 (Vals) | – | – | – | – | – | – | 55.1%#39 | Vals AI | |
| MCP Atlas | – | – | – | – | – | – | 79.2%#10 | Scale AI SEAL |
Maths
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | – | – | 46.3%#41 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | – | 17.1%#42 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | – | – | – | – | – | 90.0%#59 | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | – | – | 6.0%#53 | Vals AI |
Knowledge
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | – | – | 19.1%#64 | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience | – | – | – | – | – | – | -8.9#132 | Artificial Analysis | |
| MMLU-Pro | – | – | – | – | – | – | 85.6%#56 | Vals AI | |
| LegalBench | – | – | – | – | – | – | 83.0%#55 | Vals AI | |
| CorpFin | – | – | – | – | – | – | 69.6%#6 | Vals AI | |
| TaxEval | – | – | – | – | – | – | 75.5%#18 | Vals AI |
Human preference
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | – | – | – | – | 1407#125 | LMArena |
Multimodal
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Vision | – | – | – | – | – | – | 1234#60 | LMArena | |
| MMMU-Pro | – | – | – | – | – | – | 74.0%#88 | Artificial Analysis |
Long context
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| AA-LCR | – | – | – | – | – | – | 75.7%#107 | Artificial Analysis |
Composite
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | – | – | 150.2#42 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index | – | – | – | – | – | – | 26.1#110 | Artificial Analysis | |
| Vals Indexnot in index | – | – | – | – | – | – | 31.9#41 | Vals AI |
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 |
|---|---|---|---|---|---|---|
| Baseten | 180 tok/s | 0.27 s | $0.500 | $1.20 | 1.0M | fp8 |
| Together | 94 tok/s | 0.33 s | $0.500 | $1.20 | 524k | – |
| DeepInfra | 61 tok/s | 0.49 s | $0.450 | $1.20 | 524k | fp8 |
Price history
Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.
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 $0.100 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0005 | $0.0004 | 17.5 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0061 | $0.0030 | 19.5 s |
| Code edit | 6,000 / 1,500 | $0.0045 | $0.0029 | 25.5 s |
| Agentic coding session | 60,000 / 4,000 | $0.032 | $0.016 | 42.2 s |
| Structured extraction | 2,000 / 200 | $0.0011 | $0.0006 | 16.8 s |
See also
Data as of 9 Sept 2026. Compare these configurations.