# OpenAI vs Modal

> A model API against a serverless GPU platform with no model catalog. They do different jobs, and many teams run both: GPT for reasoning, Modal for custom models.

Canonical: https://www.subconscious.dev/compare/openai-vs-modal · By The Subconscious Team · Updated September 30, 2026

## How they compare

OpenAI and Modal rarely compete for the same line item. OpenAI sells tokens from its GPT models, priced per million, with a 1.05M window and hosted tools. Modal sells GPU seconds. A developer decorates a Python function with the hardware it needs, and Modal builds the container, autoscales it and scales it to zero, billing from $0.59 an hour for a T4 to $3.95 for an H100 at list. There is no model catalog and no per-token price, so Modal only makes sense when a team brings its own weights and serving code.

The two fit together in one product. A typical split sends reasoning and chat to GPT while Modal runs embeddings, reranking, transcription, OCR or a private fine-tune. Modal can also host OpenAI's open-weight gpt-oss for teams that want it on their own terms. Watch the production math on Modal: non-preemptible US capacity runs about 3.75x list, putting an H100 near $14.81 an hour, and keeping containers warm turns a serverless bill into an always-on one. OpenAI's equivalent trap is its 2x input rate past 272K tokens.

## What each one does

### OpenAI

OpenAI runs the most widely adopted closed-model API. Its September 2026 lineup has GPT-6 Astra at the top for computer use, coding and long agentic runs, priced at $10 in and $50 out per million tokens. Below it sits the GPT-5.6 family: Sol for hard professional work, Terra as the balanced default, and Luna for high-volume jobs at $0.20 in and $1.20 out. All of them carry a 1.05M token context window with up to 128K output.

### Modal

Modal is serverless compute with GPUs attached. A developer decorates a Python function with the hardware it needs, such as gpu="H100", and Modal builds the container, schedules it, autoscales it and scales it back to zero. Billing runs per second with no minimum increment, from $0.59 an hour for a T4 to $3.95 for an H100 at list. Containers can hold up to 8 GPUs across T4 through B300.

## Which is best, and when

### Choose OpenAI for

- Reasoning, chat and multi-tool agents with no infrastructure
- Teams without ML engineers to package models
- Computer use and coding on GPT-6 Astra

### Choose Modal for

- Bursty GPU jobs like embeddings, transcription and media processing
- Serving private or fine-tuned models on serverless endpoints
- Self-hosting open weights such as gpt-oss with per-second billing

## At a glance

| Attribute | OpenAI | Modal |
|---|---|---|
| Model access | Closed, plus open gpt-oss | Bring your own weights |
| Flagship models | GPT-6 Astra, GPT-5.6 Sol, Terra, Luna | None hosted |
| Speed | Fast mode: up to 2.5x at 2x price | ~1s container boot |
| Price | $0.20–$10 in, $1.20–$50 out per 1M | Per second; H100 $3.95/hr list |
| Customization | N/A | Run any training code |
| Deployment | API, Azure OpenAI, Bedrock | Serverless GPU containers |
| Long context | 1.05M; 2x input past 272K | Depends on the model you deploy |

## FAQ

### What is the difference between OpenAI and Modal?

A model API against a serverless GPU platform with no model catalog. They do different jobs, and many teams run both: GPT for reasoning, Modal for custom models.

### When should I choose OpenAI over Modal?

Reasoning, chat and multi-tool agents with no infrastructure; Teams without ML engineers to package models; Computer use and coding on GPT-6 Astra.

### When should I choose Modal over OpenAI?

Bursty GPU jobs like embeddings, transcription and media processing; Serving private or fine-tuned models on serverless endpoints; Self-hosting open weights such as gpt-oss with per-second billing.

### Is OpenAI or Modal cheaper?

OpenAI: $0.20–$10 in, $1.20–$50 out per 1M. Modal: Per second; H100 $3.95/hr list. The cheaper choice depends on the model and workload.

### Which has more context, OpenAI or Modal?

OpenAI: 1.05M; 2x input past 272K. Modal: Depends on the model you deploy.

## Running long-horizon agents?

If your agents run past 200K tokens, compare both against Subconscious: [Subconscious vs OpenAI](https://www.subconscious.dev/compare/subconscious-vs-openai.md), [Subconscious vs Modal](https://www.subconscious.dev/compare/subconscious-vs-modal.md).

Full profiles: [OpenAI](https://www.subconscious.dev/providers/openai.md), [Modal](https://www.subconscious.dev/providers/modal.md).