Inference.net vs Boundless

Inference.net distills your traces into custom models and runs cheap batch. Boundless keeps existing open models and tunes how they are served.

By The Subconscious Team · Updated October 29, 2026

Inference.net vs Boundless: key differences

Inference.net runs batch on idle GPU capacity, up to 1M requests per file with windows of 24 hours to 7 days. Its gateway routes open, closed and custom models under one key, captures traffic as eval and training data, and fine-tunes a task-specific model to deploy on a dedicated GPU with a 99.99% uptime target. Dedicated B200s cost $9.98 an hour. Boundless also puts open models and resold GPT and Claude models behind one key, with OpenAI and Anthropic SDK support and zero data retention. It offers no fine-tuning. On reserved capacity its engineers pick an open model on the customer's evals, tune precision, batching, caching and speculative decoding, and run it against contract targets for uptime, throughput and latency.

Inference.net is the pick for large offline jobs and for replacing a narrow closed-model workload with a smaller fine-tuned model. Boundless fits teams that would rather keep an off-the-shelf open model and have engineers tune how it is served, after a free two-week trial against their current provider.

What Inference.net and Boundless do

Inference.net

Inference.net started as a buyer of last resort for idle GPU time. Its scheduler aggregates small unused chunks of capacity across data centers and runs models on them, and it passes the steep discounts it gets from those data centers on to customers. That origin still shows in its OpenAI compatible Batch API, which takes up to 1M requests per file with completion windows from 24 hours to 7 days and far higher headroom than synchronous limits.

Example models: GLM-5.2, Kimi K2.6, Schematron, customer fine-tunes

Full Inference.net profile

Boundless

Boundless launched in September 2026 as an inference partner for AI-native startups. It sells two things. The Model API serves pre-tuned models on shared capacity, billed per token from prepaid credits with a budget per API key. Tailored inference reserves capacity for one customer, and Boundless engineers pick the model on the customer’s own evals, tune precision, batching, caching, speculative decoding and parallelism, then run the deployment against uptime, throughput and latency targets written into the contract. A free two-week evaluation tests those targets on real traffic against the customer’s current provider first.

Example models: GLM-5.3, DeepSeek V4.1 Flash, Kimi K3

Full Boundless profile

Should you choose Inference.net or Boundless?

Inference.net

Choose Inference.net for

Large offline batch jobs
Fine-tuning a model from production traces
Published dedicated GPU pricing

Boundless

Choose Boundless for

Serving-stack tuning without training
Contract targets for uptime and latency
Zero data retention

Inference.net vs Boundless at a glance

Attribute	Inference.net	Boundless
Model access	Open, closed and custom	Open weights, plus closed GPT and Claude
Flagship models	Schematron, ClipTagger, customer fine-tunes	GLM-5.3, DeepSeek V4.1 Flash, Kimi K3
Speed	Batch windows of 24h to 7 days	342 tok/s on DeepSeek V4.1 Flash 1017 (Vercel)
Price	Free pay-as-you-go; Growth $250/mo; B200 $9.98/hr	$0.04–$2.30 in, $0.21–$11.40 out per 1M (open)
Customization	Distill traces into custom models	Serving-stack tuning, no fine-tuning
Deployment	Batch API, gateway, dedicated GPUs	Serverless API, reserved tailored capacity
Long context	Varies by model	1M context on most models

Frequently asked questions

What is the difference between Inference.net and Boundless?

Inference.net distills your traces into custom models and runs cheap batch. Boundless keeps existing open models and tunes how they are served.

When should I choose Inference.net over Boundless?

Large offline batch jobs; Fine-tuning a model from production traces; Published dedicated GPU pricing.

When should I choose Boundless over Inference.net?

Serving-stack tuning without training; Contract targets for uptime and latency; Zero data retention.

Is Inference.net or Boundless cheaper?

Inference.net: Free pay-as-you-go; Growth $250/mo; B200 $9.98/hr. Boundless: $0.04–$2.30 in, $0.21–$11.40 out per 1M (open). The cheaper choice depends on the model and workload.

Which has more context, Inference.net or Boundless?

Inference.net: Varies by model. Boundless: 1M context on most models.

Related comparisons

Running long-horizon agents?

If your agents run past 200K tokens, compare both against Subconscious. Our inference stack treats a long-horizon trace as the primary workload, so speed, cost, and accuracy hold up deep into the trace.

Subconscious vs Inference.net Subconscious vs Boundless