vs
Runware vs Dolphin
Runware sells low-cost media generation on its own hardware. Dolphin sells low-cost text inference on a decentralized network of consumer GPUs.
By The Subconscious Team · Updated
Runware vs Dolphin: key differences
Runware serves 300+ priced image, video, audio and 3D models through one API on its Sonic Inference Engine, and rents raw GPUs by the second. Dolphin launched its API in October 2026 with one text model, DeepSeek V4.1 Flash, at $0.01 in and $0.07 out, served on idle consumer GPUs with no published speed, context or SLA.
There is little overlap. Runware fits media generation at volume. Dolphin fits cheap text calls where latency matters little and products that need uncensored models.
What Runware and Dolphin do
Runware
Runware sells what it calls the lowest-cost API for media generation, and it claims more than 1M developers. One endpoint covers image, video, audio, 3D and text. Every request is a task with the same shape, so switching from a Kling video to a Seedream image mostly means changing the model ID. The published rate sheet lists 300+ priced models, with images from fractions of a cent to a few cents each and video billed per second, like Seedance 2.5 at about $0.10 a second at 480p.
Example models: Seedance 2.5, Qwen-Image-3.0
Full Runware profile
Dolphin
Dolphin began in 2023 as Eric Hartford's series of open-source, uncensored fine-tunes, trained to drop the refusals and baked-in guidelines of their base models so that the developer sets the rules through the system prompt. Current releases include Dolphin Llama 3.1 8B and Dolphin Mistral 24B Venice Edition, built with Venice, and the project publishes nearly 100 models on Hugging Face. It also runs a chat app, a Telegram bot and a synthetic data service that generates answers to commonly refused prompts.
Example models: DeepSeek V4.1 Flash, Dolphin Mistral 24B
Full Dolphin profileShould you choose Runware or Dolphin?
Runware vs Dolphin at a glance
| Attribute | Runware |  Dolphin |
|---|---|---|
| Model access | Hosted media models | Open weights, uncensored fine-tunes |
| Flagship models | Seedance 2.5, Qwen-Image-3.0 | DeepSeek V4.1 Flash, Dolphin Mistral 24B |
| Speed | Unknown | Unknown |
| Price | Images from fractions of a cent | $0.01 in, $0.07 out (DeepSeek V4.1 Flash) |
| Customization | Fine-tuned diffusion checkpoints | Own uncensored fine-tunes |
| Deployment | Unified API, serverless GPUs, bare metal | Decentralized serverless API |
| Long context | Not applicable | Unknown |
Frequently asked questions
What is the difference between Runware and Dolphin?
Runware sells low-cost media generation on its own hardware. Dolphin sells low-cost text inference on a decentralized network of consumer GPUs.
When should I choose Runware over Dolphin?
Cheap image and video generation; One schema across media modalities; Raw GPUs billed by the second.
When should I choose Dolphin over Runware?
Very cheap DeepSeek V4.1 Flash calls; Models without vendor-imposed refusals; Latency-tolerant background jobs.
Is Runware or Dolphin cheaper?
Runware: Images from fractions of a cent. Dolphin: $0.01 in, $0.07 out (DeepSeek V4.1 Flash). The cheaper choice depends on the model and workload.
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.