vs
Wafer vs Dolphin
Wafer sells tuned inference on open models, billed per token. Dolphin sells one open model at a very low per-token price on consumer GPUs.
By The Subconscious Team · Updated
Wafer vs Dolphin: key differences
Wafer runs agent-tuned stacks on NVIDIA and AMD, reports 2x to 2.8x speedups over stock vLLM or SGLang, and offers per-token serverless access plus dedicated deployments. Dolphin's API launched in October 2026 with one model, DeepSeek V4.1 Flash, at $0.01 in and $0.07 out per million tokens, served on idle consumer GPUs across a decentralized network. It publishes no speed, context or SLA, and signup runs through a crypto wallet and Telegram.
Wafer is built for interactive speed, especially in coding agents. Dolphin is built for low price, and fits work where tail latency matters little.
What Wafer and Dolphin do
Wafer
Wafer builds AI agents that act as GPU performance engineers, then sells inference on the stacks those agents tune. The company came out of Y Combinator's Summer 2025 batch as a "Cursor for CUDA" that turned slow PyTorch into custom kernels. Founders Emilio Andere and Steven Arellano are based in San Francisco. Wafer raised a $4M seed led by Fifty Years in April 2026 and a $40M Series A co-led by Marathon and Chemistry in September 2026 at a $220M post-money valuation, with AMD Ventures and Wing joining. Its agents profile a workload, generate candidate configs across batching, decoding, quantization, engines, kernels and hardware, measure each one and deploy the winner.
Example models: GLM-5.3, DeepSeek V4.1 Flash
Full Wafer 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 Wafer or Dolphin?
Wafer vs Dolphin at a glance
| Attribute | Wafer |  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 | ~176 tok/s on GLM-5.3; 2–2.8x vs stock vLLM or SGLang (vendor) | Unknown |
| Price | Per token, prepaid credits | $0.01 in, $0.07 out (DeepSeek V4.1 Flash) |
| Customization | Agent-tuned dedicated deployments | Own uncensored fine-tunes |
| Deployment | Serverless, dedicated | Decentralized serverless API |
| Long context | Up to 1M (GLM-5.3, DeepSeek V4.1 Flash) | Unknown |
Frequently asked questions
What is the difference between Wafer and Dolphin?
Wafer sells tuned inference on open models, billed per token. Dolphin sells one open model at a very low per-token price on consumer GPUs.
When should I choose Wafer over Dolphin?
Large open models at interactive speed; Per-token access for coding agents; Dedicated endpoints tuned to an SLO.
When should I choose Dolphin over Wafer?
Very low prices on DeepSeek V4.1 Flash; Cheap background or batch calls; Models without vendor-imposed refusals.
Is Wafer or Dolphin cheaper?
Wafer: Per token, prepaid credits. 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.