# Subconscious vs Dolphin

> Both serve DeepSeek V4.1 Flash. Dolphin sells it at very low prices on idle consumer GPUs. Subconscious runs it on a runtime built for agent traces past 200K tokens.

Canonical: https://www.subconscious.dev/compare/subconscious-vs-dolphin · By The Subconscious Team · Updated October 11, 2026

## How they compare

Dolphin's API, launched in October 2026, serves DeepSeek V4.1 Flash at $0.01 in and $0.07 out per million tokens over a decentralized network of gamers' and hobbyists' GPUs, with validators checking that nodes run the right model. It publishes no speed, context limit or SLA, and signup runs through a crypto wallet and Telegram. Subconscious serves the same model on managed infrastructure, with a runtime that prunes the KV cache and preserves suffix state. It delivers 2x faster task completion, 50% to 80% lower cost than standard inference and a 5M+ effective context window.

Dolphin's list price is hard to beat for short, latency-tolerant calls, and its uncensored fine-tunes suit products that need models without vendor refusals. For long agent runs, predictable latency and uptime matter more than per-token price, and Dolphin cannot yet show either. Subconscious also offers dedicated and on-prem deployments and plugs into Claude Code, Cursor and GitHub Copilot.

## What each one does

### Subconscious

Subconscious is an MIT CSAIL spinout in Kendall Square that builds inference for long-horizon agents, the workloads where a single trace runs past 200K tokens and often into the millions. Its runtime drops in as a replacement for vLLM or SGLang. Instead of rereading an ever-growing context on every step, it prunes the KV cache and preserves suffix state, and Subconscious co-designs the runtime with post-trained model variants it calls Marathon. Against open models on standard inference, Subconscious delivers 2x faster task completion, delivers a 5M+ effective context window, cuts cost 50% and up to 80%, and scores neutral to 10% better on agentic benchmarks.

### 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.

## Which is best, and when

### Choose Subconscious for

- Long coding agents on DeepSeek V4.1 Flash
- 5M+ effective context
- Dedicated or on-prem deployment with standard procurement

### Choose Dolphin for

- The lowest per-token price on DeepSeek V4.1 Flash
- Uncensored open models
- Latency-tolerant batch jobs

## At a glance

| Attribute | Subconscious | Dolphin |
|---|---|---|
| Model access | Open weights | Open weights, uncensored fine-tunes |
| Flagship models | GLM 5.3, DeepSeek V4.1 Flash | DeepSeek V4.1 Flash, Dolphin Mistral 24B |
| Speed | 2x faster task completion | - |
| Price | 50–80% lower cost; billed on processed tokens | $0.01 in, $0.07 out (DeepSeek V4.1 Flash) |
| Customization | Marathon post-trained variants | Own uncensored fine-tunes |
| Deployment | Managed API, dedicated, on-prem | Decentralized serverless API |
| Long context | 5M+ effective context | - |

## FAQ

### What is the difference between Subconscious and Dolphin?

Both serve DeepSeek V4.1 Flash. Dolphin sells it at very low prices on idle consumer GPUs. Subconscious runs it on a runtime built for agent traces past 200K tokens.

### When should I choose Subconscious over Dolphin?

Long coding agents on DeepSeek V4.1 Flash; 5M+ effective context; Dedicated or on-prem deployment with standard procurement.

### When should I choose Dolphin over Subconscious?

The lowest per-token price on DeepSeek V4.1 Flash; Uncensored open models; Latency-tolerant batch jobs.

### Is Subconscious or Dolphin cheaper?

Subconscious: 50–80% lower cost; billed on processed tokens. Dolphin: $0.01 in, $0.07 out (DeepSeek V4.1 Flash). The cheaper choice depends on the model and workload.

## Try Subconscious

Subconscious speaks the OpenAI and Anthropic API formats. Base URL: https://api.subconscious.dev/v1. Docs: https://docs.subconscious.dev. Get an API key: https://platform.subconscious.dev/signin. Agent guide: https://www.subconscious.dev/agents.md.

Full profiles: [Subconscious](https://www.subconscious.dev/providers/subconscious.md), [Dolphin](https://www.subconscious.dev/providers/dolphin.md).