Inkling Small (Deep Infra)
inkling-small
Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio
Specification
most-agreed values
Context
524K
Max output
1M
Released
2026-07-30
Knowledge cutoff
—
Retires
2026-10-16
Open weights
yes
Input
text, image, audio
Output
text
Price
US dollars per million tokens · most-agreed
Input
$0.45
Output
$1.20
Cache read
$0.10
Quality
6 benchmarks · 1 source
Composite
48th
percentile of 344 scored models
Rank
179 / 344
±0.06 sd
Evidence
6 × 1
one source, or fewer than 3 benchmarks
Effort range
—
one configuration only
reasoning59th2/4 bench
math60th3/7 bench
knowledge12th1/1 bench
Every score
one row per source and configuration — nothing averaged away
| Benchmark | Score | Configuration | Source | Run |
|---|---|---|---|---|
| GPQA diamond | 88.5 ±1.9 | xhigh | Epoch AI | 2026-08-14 |
| OTIS Mock AIME 2024-2025 | 90.0 ±4.0 | xhigh | Epoch AI | 2026-08-14 |
| FrontierMath-Tiers-1-3-v2 | 46.3 ±3.0 | xhigh | Epoch AI | 2026-08-14 |
| FrontierMath-Tier-4-v2 | 17.1 ±5.9 | xhigh | Epoch AI | 2026-08-14 |
| Mystery Game Puzzles | 6.0 ±2.4 | xhigh | Epoch AI | 2026-08-15 |
| SimpleQA Verified | 19.1 ±1.2 | xhigh | Epoch AI | 2026-08-27 |
The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.
Available from 17 hosts
| Host | In $/M | Out $/M | Cache rd | Cache wr | Context | Output | Retires |
|---|---|---|---|---|---|---|---|
| openrouter · free | free | free | — | — | 1M | 262K | — |
| thinkingmachines · peft:262144:sampling-nvfp4 | $0.30 | $1.20 | $0.06 | — | 262K | 262K | — |
| deepinfra | $0.45 | $1.20 | $0.10 | — | 524K | 524K | — |
| deepinfra · thinkingmachines | $0.45 | $1.20 | $0.10 | — | 524K | 1M | — |
| edenai | $0.45 | $1.20 | $0.10 | — | 524K | 1M | — |
| kilo · thinkingmachines | $0.45 | $1.20 | $0.10 | — | 524K | 262K | — |
| nano-gpt · thinking | $0.45 | $1.20 | $0.10 | — | 524K | 33K | — |
| nano-gpt · thinkingmachines | $0.45 | $1.20 | $0.10 | — | 524K | 33K | — |
| openrouter · thinkingmachines | $0.45 | $1.20 | $0.10 | — | 524K | 262K | — |
| vercel · thinkingmachines | $0.45 | $1.20 | $0.10 | — | 1M | 1M | — |
| arcee · thinkingmachines | $0.50 | $1.20 | $0.10 | — | 262K | 262K | — |
| baseten | $0.50 | $1.20 | $0.10 | — | 1M | 33K | 2026-10-16 |
| commandcode · thinkingmachines | $0.50 | $1.20 | $0.10 | — | 1M | 1M | — |
| huggingface · thinkingmachines | $0.50 | $1.20 | — | — | 524K | 1M | — |
| llmtr · thinkingmachines | $0.58 | $1.44 | $0.116 | — | 262K | 262K | — |
| thinkingmachines · thinkingmachines | $0.58 | $1.44 | $0.116 | — | 66K | 66K | — |
| aimlapi · thinkingmachines | $0.6136 | $1.5235 | $0.1227 | — | 524K | 262K | — |
Price history
append-only observations · a listing writes a row only when its price moves
| Date | Host | Field | $/M |
|---|---|---|---|
| 2026-09-02 | edenai | Input | $0.45 |
| 2026-09-18 | edenai | Input | $0.45 |
| 2026-09-02 | nano-gpt · thinking | Input | $0.50 |
| 2026-10-19 | nano-gpt · thinking | Input | $0.45 |
| 2026-09-02 | nano-gpt · thinkingmachines | Input | $0.50 |
| 2026-10-19 | nano-gpt · thinkingmachines | Input | $0.45 |
| 2026-09-02 | openrouter · free | Input | free |
| 2026-09-19 | openrouter · free | Input | free |
| 2026-10-10 | openrouter · free | Input | free |
| 2026-09-02 | vercel · thinkingmachines | Input | $0.50 |
| 2026-09-28 | vercel · thinkingmachines | Input | $0.45 |
| 2026-09-02 | openrouter · free | Output | free |
| 2026-09-19 | openrouter · free | Output | free |
| 2026-10-10 | openrouter · free | Output | free |
| 2026-10-13 | aimlapi · thinkingmachines | Cache read | $0.1595 |
| 2026-10-23 | aimlapi · thinkingmachines | Cache read | $0.1227 |
| 2026-10-13 | aimlapi · thinkingmachines | Input | $0.7977 |
| 2026-10-23 | aimlapi · thinkingmachines | Input | $0.6136 |
| 2026-10-13 | aimlapi · thinkingmachines | Output | $1.9806 |
| 2026-10-23 | aimlapi · thinkingmachines | Output | $1.5235 |