Thinking Machines: Inkling 2 Mini
thinkingmachines/inkling-2-mini
Inkling 2 Mini is an open-weight multimodal looped transformer from Thinking Machines Lab, with 3.4B active parameters out of 26.4B total (MoE). Its middle layers run three times with shared weights, so it does more computation per token than its size suggests while staying small enough to run on a laptop or a single workstation GPU. It is the smallest model in the second Inkling generation, below Inkling 2 SmallOpens in new tab, and is suited for agentic and tool-use systems, coding assistants, chatbots and retrieval-augmented generation, with English and general multilingual support.
Providers
This model is hosted by one provider. OpenRouter forwards every request without routing.
| $0.091 | $0.362 | $0.018 | 1.30s | 56 tok/s |
- Input /M
- $0.091
- Output /M
- $0.362
- Cache read /M
- $0.018
- Latency
- 1.30s
- Throughput
- 56 tok/s
- Uptime
- 100.00%
Pricing
The average price customers actually pay for this model, next to the prices providers post. Caching and discounts mean the price actually paid is often well below the listed one.
Weighted average
Inkling 2 Mini — Price history
| $0.0359 | $0.388 | 78.5% | 100.0% |
- Effective in $0.0359 / Effective out $0.388Token share100.0%
Performance
Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better).
P50, best across providers
P50, best provider
Uptime
Uptime is the percentage of the past 3 days that at least one provider was responding to requests. Availability is the percentage of time that inference was successfully served. OpenRouter continuously monitors and uses the next-best provider when one returns an error.
Availability over the last 3 days
Last 72 hoursAvailability over the last 24 hours
When an error occurs in an upstream provider, we can recover by routing to another healthy provider, if your request filters allow it. You can access per-provider uptime data programmatically through the Endpoints API. Learn more about our load balancing and provider routing options.
Benchmarks
Scores on standardized evaluations and this model’s own outputs on shared benchmark prompts. Higher percentages are better — and each rank shows where this model lands among the models evaluated on that source.
| Source | Benchmark | Score |
|---|---|---|
| Artificial Analysis | Inkling 2 Mini Intelligence Index | 20.1 |
| Artificial Analysis | Inkling 2 Mini Finance & Accounting Index | 35.3 |
| Artificial Analysis | Inkling 2 Mini Strategy & Ops Index | 35.2 |
| Artificial Analysis | Inkling 2 Mini Legal Index | 31.0 |
| Artificial Analysis | Inkling 2 Mini Engineering Index | 27.6 |
| Artificial Analysis | Inkling 2 Mini Economics Index | 33.0 |
| Artificial Analysis | Inkling 2 Mini AutomationBench-AA | 41.8% |
| Artificial Analysis | Inkling 2 Mini GDP.pdf | 16.8% |
| Artificial Analysis | Inkling 2 Mini AA-AnalystAgent | 46.3% |
| Artificial Analysis | Inkling 2 Mini EnterpriseOps-Gym-AA | 42.0% |
| Artificial Analysis | Inkling 2 Mini GDPval-AA | 40.9% |
| Artificial Analysis | Inkling 2 Mini HLE-Diamond | 15.2% |
| Artificial Analysis | Inkling 2 Mini Terminal-Bench 4.0 | 5.6% |
| Artificial Analysis | Inkling 2 Mini FrontierCode-AA 2.0 | 15.1% |
| Artificial Analysis | Inkling 2 Mini Terminal-Bench-Science | 1.4% |
| Artificial Analysis | Inkling 2 Mini CritPt v2 (Hard Subset) | 17.0% |
| Artificial Analysis | Inkling 2 Mini CritPt v2 | 31.4% |
| Artificial Analysis | Inkling 2 Mini AA-Omniscience Accuracy | 20.5% |
| Artificial Analysis | Inkling 2 Mini AA-Omniscience Non-Hallucination Rate | 63.4% |
Apps
Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.
Activity
Token volume and request traffic to this model over time.
API
Drop-in code to call this model. OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. The only thing that changes between models is the model slug below.
Not currently available through EU in-region routing. In-region routing docs
Frequently asked questions
Inkling 2 Mini is an open-weight multimodal looped transformer from Thinking Machines Lab, with 3.4B active parameters out of 26.4B total (MoE). Its middle layers run three times with shared weights, so it does more computation per token than its size suggests while staying small enough to run on a laptop or a single workstation GPU.
Inkling 2 Mini costs $0.091/M input tokens and $0.362/M output tokens, with separate rates for Cache Read at $0.018/M tokens.
Inkling 2 Mini has a 1,048,576 token context window. It supports up to 65,536 completion tokens.
Yes. Inkling 2 Mini accepts tools and tool_choice for function calling. It also supports structured outputs via a JSON schema in response_format.
Inkling 2 Mini accepts text and images as input and returns text.
Inkling 2 Mini is served by 1 provider on OpenRouter: Prime Intellect.
Inkling 2 Small, Inkling Small, and Inkling are other text models from Thinking Machines.
Inkling 2 Mini was released on November 21, 2026.