Wafer vs General Compute

Both sell speed on open models. Wafer tunes the GPU serving stack, and General Compute moves decode onto Cerebras, SambaNova and other ASICs.

By The Subconscious Team · Updated October 27, 2026

Wafer vs General Compute: key differences

Wafer's agents tune batching, kernels and engines per deployment on NVIDIA or AMD. Artificial Analysis measures it at about 176 tokens per second on GLM-5.3 and 152 on DeepSeek V4.1 Flash, both with up to 1M context, and it offers OpenAI- and Anthropic-compatible APIs billed per token. General Compute runs prefill on NVIDIA GPUs and decode on ASICs, and in its own benchmark measured over 1,000 tokens per second on GPT-OSS 120B. Its catalog is older open models, from DeepSeek V3.2 with 32K context to MiniMax M2.7 with 192K, at $0.21 to $0.28 in per million tokens.

Pick Wafer for newer large open models with 1M context, an Anthropic-compatible endpoint for coding tools, or a dedicated endpoint that keeps getting re-tuned as load changes. Pick General Compute when raw decode speed on its catalog matters most, or when you want reserved racks with a tokens-per-second SLA.

What Wafer and General Compute 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

General Compute

General Compute is a San Francisco inference cloud built on alternative chips. It splits each request across hardware, with NVIDIA GPUs handling prefill and ASICs from Cerebras, SambaNova, Positron AI and d-Matrix handling decode. In its own open benchmark on GPT-OSS 120B it measured 1,025 to 1,234 output tokens per second against 100 to 121 on an unnamed GPU cloud, and up to 16.2x faster end to end. Founders Finn Puklowski and Jason Goodison raised a $15M seed led by FUSE, an up to $400M debt facility from Upper90 and, on October 27, 2026, a $100M Series A co-led by Blackbird and Abstract.

Example models: MiniMax M2.7, DeepSeek V3.2, GPT-OSS 120B

Full General Compute profile

Should you choose Wafer or General Compute?

Wafer

Choose Wafer for

Newer open models with 1M context
Anthropic-compatible API for coding tools
Dedicated endpoints on NVIDIA or AMD

General Compute

Choose General Compute for

High decode speed on its catalog
Bring-your-own LoRA checkpoints
Tokens-per-second SLAs

Wafer vs General Compute at a glance

Attribute	Wafer	General Compute
Model access	Open weights	Open weights
Flagship models	GLM-5.3, DeepSeek V4.1 Flash	MiniMax M2.7, DeepSeek V3.2, GPT-OSS 120B
Speed	~176 tok/s on GLM-5.3; 2–2.8x vs stock vLLM or SGLang (vendor)	1,000+ tok/s on GPT-OSS 120B (vendor)
Price	Per token, prepaid credits	$0.21–$0.28 in, $0.38–$1.20 out per 1M
Customization	Agent-tuned dedicated deployments	Custom LoRA and fine-tuned checkpoints
Deployment	Serverless, dedicated	Serverless API, dedicated racks, bare metal
Long context	Up to 1M (GLM-5.3, DeepSeek V4.1 Flash)	192K (MiniMax M2.7)

Frequently asked questions

What is the difference between Wafer and General Compute?

Both sell speed on open models. Wafer tunes the GPU serving stack, and General Compute moves decode onto Cerebras, SambaNova and other ASICs.

When should I choose Wafer over General Compute?

Newer open models with 1M context; Anthropic-compatible API for coding tools; Dedicated endpoints on NVIDIA or AMD.

When should I choose General Compute over Wafer?

High decode speed on its catalog; Bring-your-own LoRA checkpoints; Tokens-per-second SLAs.

Is Wafer or General Compute cheaper?

Wafer: Per token, prepaid credits. General Compute: $0.21–$0.28 in, $0.38–$1.20 out per 1M. The cheaper choice depends on the model and workload.

Which has more context, Wafer or General Compute?

Wafer: Up to 1M (GLM-5.3, DeepSeek V4.1 Flash). General Compute: 192K (MiniMax M2.7).

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.

Subconscious vs Wafer Subconscious vs General Compute