Thinking Machines vs Inception

Thinking Machines sells post-training for open models through Tinker. Inception sells fast closed models with fine-tuning only on request.

By The Subconscious Team · Updated October 11, 2026

Thinking Machines vs Inception: key differences

Thinking Machines' Tinker lets teams write their own SFT and RL loops with LoRA on models like Kimi K2.6, Qwen3.5 and its own Apache 2.0 Inkling models, which take text, image and audio with up to 1M context. A beta serverless API serves only Inkling, at $1.00 in and $4.05 out. 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.

These solve different problems. Thinking Machines is for teams training their own model on open weights. Inception is for teams that want a fast, cheap hosted model ready for production traffic now.

What Thinking Machines and Inception do

Thinking Machines

Thinking Machines Lab is the San Francisco lab Mira Murati, formerly CTO of OpenAI, founded in February 2025, with OpenAI co-founder John Schulman as chief scientist. It raised about $2 billion at a $12 billion valuation in July 2025 in a round led by Andreessen Horowitz, and in March 2026 signed a multi-year Nvidia deal for one gigawatt of Vera Rubin capacity. In September 2026 it signed a $65M-a-year deal with Crusoe to serve Inkling, GLM 5.2 and 5.3 and its fine-tunes on dedicated B200s. Its main developer product is Tinker, an API for post-training open-weight models that launched in October 2025 and is now generally available. Tinker exposes four low-level calls, forward_backward, optim_step, sample and save_state, so teams write their own supervised or reinforcement learning loops while Thinking Machines runs the distributed GPU work. Training uses LoRA adapters rather than full weight updates.

Example models: Inkling, Inkling-Small, Qwen3.5, Kimi K2.6

Full Thinking Machines profile

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.

Example models: Mercury 2.5, Mercury 2

Full Inception profile

Should you choose Thinking Machines or Inception?

Thinking Machines

Choose Thinking Machines for

Custom SFT and RL on open models
Fine-tuning large MoE models
Inkling with image, audio and 1M context

Inception

Choose Inception for

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

Thinking Machines vs Inception at a glance

Attribute	Thinking Machines	Inception
Model access	Open weights	Closed weights (Mercury dLLMs)
Flagship models	Inkling, Inkling-Small	Mercury 2.5, Mercury 2
Speed	Unknown	1,100+ tok/s (vendor); ~640 measured
Price	Per 1M tokens by prefill, sample, train; Inkling $1.00 in, $4.05 out	$0.20 in, $0.75 out (Mercury 2.5)
Customization	LoRA SFT and RL via Tinker	Fine-tuning on request
Deployment	Training API, beta serverless (Inkling only)	Serverless API, Bedrock, Azure, private
Long context	Inkling up to 1M (256K serverless); Tinker 32K–256K	260K

Frequently asked questions

What is the difference between Thinking Machines and Inception?

Thinking Machines sells post-training for open models through Tinker. Inception sells fast closed models with fine-tuning only on request.

When should I choose Thinking Machines over Inception?

Custom SFT and RL on open models; Fine-tuning large MoE models; Inkling with image, audio and 1M context.

When should I choose Inception over Thinking Machines?

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

Is Thinking Machines or Inception cheaper?

Thinking Machines: Per 1M tokens by prefill, sample, train; Inkling $1.00 in, $4.05 out. Inception: $0.20 in, $0.75 out (Mercury 2.5). The cheaper choice depends on the model and workload.

Which has more context, Thinking Machines or Inception?

Thinking Machines: Inkling up to 1M (256K serverless); Tinker 32K–256K. Inception: 260K.

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 Thinking Machines Subconscious vs Inception