Inkling
Thinking MachinesOpen weightsApache 2.0Released Jul 15, 2026
On the arena.ai Agents leaderboard, Inkling has a Net Improvement of -17.32% (—, rank #57, 95% interval -18.88%–-15.76%). Inkling ranks #95 on the LMArena Text leaderboard (style control) with an Arena score of 1442.5 (±4.1, 50,718 votes). Artificial Analysis Intelligence Index: 12.5. API price: $1.00 input / $4.05 output per 1M tokens. Context: 524K. Nearest rivals: Grok 4.7 (1443.3) above and Qwen3.5 397B (1441.4) below.
Quick answer
- Agents leaderboard
- #57 of 56 · -17.32% success
- Text (LMArena)
- #95 · 1442.5
- 17NAS overall
- #53 of 55
- API price (in / out)
- $1.00 / $4.05 per 1M tokens
- Closest rival
- Grok 4.7 · 1443.3
- Cheaper alternative
- Qwen3.7 Plus: Text 1455.3, output $1.60 (60% cheaper)
- Data updated
- Nov 15, 2026 · synced daily
Sources: arena.ai, LMArena, Artificial Analysis, OpenRouter
Key metrics
Models ranked nearby
Sorted by Arena score| # | Model | Arena | Intel. | Output $ | tok/s |
|---|---|---|---|---|---|
| 58 |  MiMo V2.6 Flash | 1448.4 | 26.7 | $0.28 | 34 |
| 59 |  Gemini 2.5 Pro | 1445.5 | – | $10.00 | 116 |
| 60 |  Grok 4.7 | 1443.3 | 35.5 | $6.00 | 74 |
| 61 | TInklingThis page | 1442.5 | 12.5 | $4.05 | 178 |
| 62 |  Qwen3.5 397B | 1441.4 | – | $3.60 | – |
| 63 |  GPT-6 Luna | 1440.9 | 27.9 | $0.50 | 123 |
| 64 | GPT-5.2 | 1437.2 | – | $14.00 | 78 |
More from Thinking Machines
| # | Model | Arena | Intel. | Output $ | tok/s |
|---|---|---|---|---|---|
| 75 | TInkling 2 Small | 1411.8 | 30.6 | $1.00 | 103 |
| 76 | TInkling Small | 1405.4 | 12.9 | $1.20 | 199 |
Pricing
- Official API list price: $1.00 input and $4.05 output per 1M tokens (USD).
- At a typical 3:1 input-to-output ratio, the blended cost is about $1.76 per 1M tokens.
- Among 95 models with an output price, it is the #49 cheapest.
- The cheapest OpenRouter route charges $1.00 input / $4.05 output.
Sources & methodology
On the arena.ai Agents leaderboard, Inkling has a Net Improvement of -17.32% (—, rank #57, 95% interval -18.88%–-15.76%).
- arena.ai Agents — agent Net Improvement, 95% interval, sessions and rank (via the public Hugging Face dataset, CC BY 4.0).
- LMArena — Arena score (Text leaderboard with style control, the same view arena.ai shows by default), 95% confidence interval, votes and rank.
- Artificial Analysis — Intelligence / Coding Index, output speed and latency (medians).
- OpenRouter — Route prices, context length and listing date.
Arena scores use the Bradley-Terry model (not Elo); ± is the half-width of the 95% confidence interval. Latency figures are Artificial Analysis medians; for reasoning models, time to first token and to first answer include the thinking phase. Data syncs daily at 14:20 (UTC+8).
Limits: Arena scores come from anonymous users voting on chats, so they reflect conversational preference rather than agent or coding benchmarks; with few votes the interval is wide and ranks move. Prices and speeds change as providers update them.
FAQ
How does Inkling do on the Agents leaderboard?
On the arena.ai Agents leaderboard, Inkling has a Net Improvement of -17.32% (—, rank #57, 95% interval -18.88%–-15.76%).
What rank is Inkling right now?
Inkling is #61 on our leaderboard of 80 scored models (Arena score 1442.5); its raw LMArena rank is #95.
What does Inkling cost via API?
Input $1.00 / output $4.05 per 1M tokens (official API list price; the OpenRouter quote when no official price is available). At a typical 3:1 input-to-output ratio, the blended cost is about $1.76 per 1M tokens.
How fast is Inkling?
Artificial Analysis measures about 178 output tokens per second, 2.2 s to the first token and 13 s to the first answer token; it is a reasoning model, so latency includes thinking time.
How close is Inkling to the models ranked next to it?
It is 0.8 points behind Grok 4.7 (1443.3) and 1.1 points ahead of Qwen3.5 397B (1441.4); gaps smaller than the confidence interval (±4.1) are not significant.
Is Inkling open source? Can it be self-hosted?
Inkling is an open-weights model (license: Apache 2.0); you can download the weights and self-host.
What is the context length of Inkling?
524K (524,288 tokens, as listed on OpenRouter).
Inkling vs Grok 4.7: which is better?
On the LMArena Text leaderboard Inkling scores 1442.5 and Grok 4.7 scores 1443.3 (a 0.8-point gap, within the confidence interval, so not significant). Agent task success: Inkling -17.32%, Grok 4.7 2.56%. On price, Inkling costs $4.05 and Grok 4.7 $6.00 per 1M output tokens.
Is there a cheaper model close to Inkling?
Yes: Qwen3.7 Plus scores 1455.3 on the Text leaderboard (Inkling: 1442.5) and costs $1.60 per 1M output tokens, about 60% less than Inkling ($4.05).
Where does the Inkling data come from and how often is it updated?
Agent success rates come from the arena.ai Agents leaderboard (CC BY 4.0), Text scores from LMArena, intelligence and speed from Artificial Analysis, and prices and context from official list prices and OpenRouter. 17NAS syncs daily at 06:20 UTC; this page was last updated Nov 15, 2026.