vs
Anthropic vs Inception
Anthropic sells frontier Claude models known for agentic coding. Inception sells diffusion LLMs that trade peak intelligence for very fast output.
By The Subconscious Team · Updated
Anthropic vs Inception: key differences
Claude runs from Fable 5.1 at $10 in and $50 out down to Haiku 5.5 at $0.10 in and $0.50 out, with 1M context on every tier and $0.25 cache reads on Fable. Fable always thinks, so it is the slowest tier. Inception's Mercury 2.5 is a closed reasoning model with 260K context, tool use and JSON mode at $0.20 in and $0.75 out per million. Inception reports over 1,100 tokens per second, and Artificial Analysis measured about 640. Its quality sits near Claude Haiku 4.5, not Fable or Opus.
Anthropic is the pick for long agent runs, hard coding tasks and code review, where model quality decides the outcome. Inception fits the latency-critical steps around them, like autocomplete, quick edits and voice, with much faster output than Claude. Both are sold through AWS Bedrock, so a team can use the two inside one cloud account.
What Anthropic and Inception do
Anthropic
Anthropic sells the Claude family of closed models through its own API, Amazon Bedrock, Google Vertex AI and Microsoft Foundry. The public lineup today runs from Claude Fable 5.1 at the top, released September 1, 2026, through Opus 5.5 ($4 in, $20 out) and Sonnet 5.5 ($2 in, $10 out) down to Haiku 5.5. List prices span a hundredfold range, from $10 in and $50 out on Fable to $0.10 in and $0.50 out on Haiku. All four tiers have a 1M token context window. The top three bill it at standard pricing, while Haiku 5.5 charges more past 100K.
Example models: Claude Fable 5.1, Claude Opus 5.5, Claude Haiku 5.5
Full Anthropic 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 profileShould you choose Anthropic or Inception?
Anthropic vs Inception at a glance
| Attribute | Anthropic |  Inception |
|---|---|---|
| Model access | Closed | Closed weights (Mercury dLLMs) |
| Flagship models | Claude Fable 5.1, Opus 5.5, Sonnet 5.5, Haiku 5.5 | Mercury 2.5, Mercury 2 |
| Speed | Fable is the slowest tier | 1,100+ tok/s (vendor); ~640 measured |
| Price | $0.10–$10 in, $0.50–$50 out per 1M | $0.20 in, $0.75 out (Mercury 2.5) |
| Customization | N/A | Fine-tuning on request |
| Deployment | API, Bedrock, Vertex AI, Microsoft Foundry | Serverless API, Bedrock, Azure, private |
| Long context | 1M; Haiku 5.5 costs more past 100K | 260K |
Frequently asked questions
What is the difference between Anthropic and Inception?
Anthropic sells frontier Claude models known for agentic coding. Inception sells diffusion LLMs that trade peak intelligence for very fast output.
When should I choose Anthropic over Inception?
Agentic coding and code review; Long-running agents with 1M context; Hard reasoning where quality decides.
When should I choose Inception over Anthropic?
Autocomplete and fast code edits; Voice agents and real-time chat; Haiku-class quality at very high speed.
Is Anthropic or Inception cheaper?
Anthropic: $0.10–$10 in, $0.50–$50 out per 1M. Inception: $0.20 in, $0.75 out (Mercury 2.5). The cheaper choice depends on the model and workload.
Which has more context, Anthropic or Inception?
Anthropic: 1M; Haiku 5.5 costs more past 100K. 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.