Inkling-Smallとは？上位版との推論・コーディング性能の比較

"Today, we are releasing Inkling-Small, an efficient open-weights model that achieves comparable performance to Inkling at a quarter of its size."／"Inkling-Small is a Mixture-of-Experts transformer with 276B total parameters, 12B active, trained on NVIDIA GB300 NVL72 systems. Like Inkling, it features native reasoning over audio and images, variable thinking effort, a context window of up to 1M tokens, and well-rounded performance across a range of benchmarks."／"975B total"／"41B active" — Thinking Machines Lab 公式発表（2026年7月30日）の冒頭とパラメータ比較図より

"Inkling-Small is a general-purpose multimodal model that accepts text, image and audio inputs and generates text outputs."／"A 42-layer decoder-only transformer with a sparse Mixture-of-Experts (MoE) feed-forward backbone: each token is routed to 6 of 256 experts, plus 2 shared experts active on every token."／"276B total, 12B active"／"Inkling-Small supports a context window of up to 1M tokens." — Inkling-Small モデルカード「Intended uses」「Architecture type」「Parameters」より

"Jul 30, 2026"／"We are releasing the full weights of Inkling-Small." — Thinking Machines Lab 公式発表より

"Date of release"／"July 30, 2026"／"License"／"Apache 2.0"／"It is released with open weights to support research, fine-tuning and integration into third-party products by downstream developers."／"Model Acceptable Use Policy"／"Training Data Documentation" — Inkling-Small モデルカード「1. General information」「8. Legal」より

"Output TFLOPs per sample are estimated as 2 × active parameters × mean generated tokens per sample"／"Inkling output pricing is $4.05 / 1M tokens and Inkling-Small output pricing is $1.20 / 1M tokens" — Thinking Machines Lab 公式発表の性能と計算量・費用の比較図の説明より

"Inkling-Small began training after its larger counterpart, which let us improve its training process. For example, we made changes to Inkling-Small’s pre-training data mix and machine learning recipe."／"With these improvements, Inkling-Small surpassed Inkling on reasoning and agentic coding benchmarks." — Thinking Machines Lab 公式発表「Capabilities」節より

"we made changes to Inkling-Small’s pre-training data mix and machine learning recipe."／"we post-trained an earlier checkpoint"／"Inkling-Small (preview)"／"in part using on-policy distillation with Inkling as the teacher." — Thinking Machines Lab 公式発表「Capabilities」節より

"Starting from that checkpoint, we continued scaling agentic coding RL for two weeks." — Thinking Machines Lab 公式発表「Capabilities」節より

"On Humanity’s Last Exam it scores 31.6%, ahead of Inkling’s 29.7%, and the advantage holds at every thinking budget: Inkling-Small’s test-time compute curves sit above Inkling’s throughout. On SWEBench-Verified it exceeds 80%."／"All evals run at effort 0.99 and temperature 1.0."／"Inkling and Inkling-Small’s numbers are reported using a bash-only harness."／"Inkling and Inkling-Small’s numbers are reported using an internal coding harness. A small number of solutions were found to be contaminated from web search and were assigned a score of 0." — Thinking Machines Lab 公式発表「Reasoning and Agentic Tasks」「Benchmarking Inkling-Small」節と評価表の注記より

"Inkling maintains an advantage on knowledge coverage and factuality."／"Inkling-Small against open- and closed-weights models across the full eval suite." — Thinking Machines Lab 公式発表「Capabilities」節と評価表の説明より

"We show that Inkling-Small is competitive with other open-weights models in a similar size range on both performance and efficiency."／"Inkling-Small against open- and closed-weights models across the full eval suite. Activated and total parameters are given for scale; a dash means the score was not available at the time of writing." — Thinking Machines Lab 公式発表の性能比較図と評価表の説明より

"Furthermore, its variable thinking effort lets users easily adapt it to target their use case, balancing cost and performance."／"Sweeping reasoning effort from minimal to xhigh traces the performance-cost curve (dollar output price per sample) for Inkling-Small and Inkling on Terminal-Bench 2.1, HLE (no tool), and IFBench."／"Estimated output cost per sample is computed as mean generated tokens per sample × output price per token, where generated-token counts include reasoning tokens" — Thinking Machines Lab 公式発表の性能と費用の比較図の説明より

"The model is distributed in two checkpoint formats."／"The NVFP4 checkpoint offers a quantized alternative that reduces the aggregated VRAM requirement to at least 180 GB."／"W4A4 on 1x NVIDIA B300 GPU (note: W4A4 mode additionally requires SM100+ architecture)"／"W4A16 on 2x NVIDIA H200 GPUs" — Inkling-Small モデルカード「Required hardware」より

"The BF16 checkpoint requires a GPU cluster with at least 600 GB of aggregated VRAM. This can be met with either of the following configurations:"／"4x NVIDIA B300 GPUs"／"8x NVIDIA H200 GPUs" — Inkling-Small モデルカード「Required hardware」より

"Running the model directly on GPU hardware requires an inference deployment framework–either SGLang, vLLM, TokenSpeed, Unsloth, or Hugging Face, along with all of their respective dependency libraries." — Inkling-Small モデルカード「Required software」より

"Image: Any pixel-based image input. For optimal performance, each image dimension should be between 40px to 4096px."／"Audio: WAV format, sampled at 16kHz. For optimal performance, audio length should ideally be under 2 mins."／"Inkling-Small supports a context window of up to 1M tokens."／"For all input images whose long edge is smaller than 2048 px, we resize the long edge to the smaller of 2048 px or 2× its original length." — Inkling-Small モデルカード「Input modalities」と「5. Evaluations」の注記より

"The weights are available for download through Hugging Face"／"Developers can also access the model using the Tinker fine-tuning platform, which provides API access."／"The model is also available via API access through third party inference providers."／"For accessing our model via Tinker, you can get started by referring to the Tinker Cookbook"／"For accessing our model via third party inference providers, you can refer to the documentation from our inference provider partners." — Inkling-Small モデルカード「3. Methods of distribution」より

"We are releasing the full weights of Inkling-Small. We’re also making it available for fine-tuning on Tinker, and for text, image, and audio chat on Tinker Playground" — Thinking Machines Lab 公式発表より

"License"／"Apache 2.0"／"Inkling-Small may exhibit general limitations common to foundation models, including hallucination (generating plausible but factually incorrect or unsupported content), occasional failures to follow instructions precisely, and degraded performance in long multi-turn conversations."／"Inkling-Small may reflect biases present in its training data, including demographic, cultural, or linguistic biases, and may perform unevenly across languages, dialects, or subject domains that were less represented during training."／"Inkling-Small’s knowledge is limited to information available as of its training cutoff"／"We recommend that downstream developers and deployers apply appropriate human oversight and review for outputs used in high-stakes or safety-critical contexts"／"Avoid deploying Inkling-Small in domains such as medical, legal, or safety-critical decision-making without additional fine-tuning, domain-specific validation, and human oversight"／"Inkling-Small’s occasional tendency to comply with role-play and indirectly framed prompts concerning harmful topics"／"Common downstream moderation tools, such as Llama Guard, are compatible with Inkling-Small and can be layered around the model to catch jailbreak attempts, filter unsafe outputs, and enforce use-case-specific policies." — Inkling-Small モデルカード「1. General information」「6. Safety」「7. Bias, risks and limitations」より