# Wafer vs Inception

> Wafer makes open models faster by tuning the serving stack. Inception makes its own models faster with a diffusion architecture.

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

## How they compare

Wafer's agents tune batching, kernels and engines for each deployment on NVIDIA or AMD, and it reports 2x to 2.8x speedups over stock vLLM or SGLang on models like Qwen 3.5 397B and GLM 5.1. Its serverless API bills per token. Inception sells closed Mercury diffusion models that refine a block of tokens in parallel instead of one at a time. Mercury 2.5 is a reasoning model with tool use and 260K context at $0.20 in and $0.75 out per million tokens; Inception reports over 1,100 tokens per second, and Artificial Analysis measures about 640.

Pick Wafer for large open models at interactive speed, or a dedicated endpoint tuned to a latency SLO. Pick Inception when its closed models are good enough for the task and raw output speed is the priority.

## What each one does

### 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.

### Inception

Inception builds diffusion large language models (dLLMs). Every other LLM on this list generates text one token at a time; Inception's Mercury models start from a rough draft of a whole block and refine it over a few passes, producing many tokens in parallel. That ties speed less to output length and lets the models run fast on ordinary NVIDIA GPUs rather than custom chips. The company was founded by professors from Stanford, UCLA and Cornell behind core work on diffusion, Flash Attention and DPO, and is led by Stanford's Stefano Ermon. It raised a $50M seed in November 2025 led by Menlo Ventures, with Mayfield, M12, NVentures, Snowflake and Databricks, plus Andrew Ng and Andrej Karpathy as angels.

## Which is best, and when

### Choose Wafer for

- Large open models at interactive speed
- Per-token access for coding agents
- Dedicated endpoints tuned to an SLO

### Choose Inception for

- Very fast output on standard GPUs
- Cheap reasoning with tool use and 260K context
- Buying through AWS Bedrock or Azure

## At a glance

| Attribute | Wafer | Inception |
|---|---|---|
| Model access | Open weights | Closed weights (Mercury dLLMs) |
| Flagship models | GLM-5.3, DeepSeek V4.1 Flash | Mercury 2.5, Mercury 2 |
| Speed | ~176 tok/s on GLM-5.3; 2–2.8x vs stock vLLM or SGLang (vendor) | 1,100+ tok/s (vendor); ~640 measured |
| Price | Per token, prepaid credits | $0.20 in, $0.75 out (Mercury 2.5) |
| Customization | Agent-tuned dedicated deployments | Fine-tuning on request |
| Deployment | Serverless, dedicated | Serverless API, Bedrock, Azure, private |
| Long context | Up to 1M (GLM-5.3, DeepSeek V4.1 Flash) | 260K |

## FAQ

### What is the difference between Wafer and Inception?

Wafer makes open models faster by tuning the serving stack. Inception makes its own models faster with a diffusion architecture.

### When should I choose Wafer over Inception?

Large open models at interactive speed; Per-token access for coding agents; Dedicated endpoints tuned to an SLO.

### When should I choose Inception over Wafer?

Very fast output on standard GPUs; Cheap reasoning with tool use and 260K context; Buying through AWS Bedrock or Azure.

### Is Wafer or Inception cheaper?

Wafer: Per token, prepaid credits. Inception: $0.20 in, $0.75 out (Mercury 2.5). The cheaper choice depends on the model and workload.

### Which has more context, Wafer or Inception?

Wafer: Up to 1M (GLM-5.3, DeepSeek V4.1 Flash). Inception: 260K.

## Running long-horizon agents?

If your agents run past 200K tokens, compare both against Subconscious: [Subconscious vs Wafer](https://www.subconscious.dev/compare/subconscious-vs-wafer.md), [Subconscious vs Inception](https://www.subconscious.dev/compare/subconscious-vs-inception.md).

Full profiles: [Wafer](https://www.subconscious.dev/providers/wafer.md), [Inception](https://www.subconscious.dev/providers/inception.md).