Inception: Mercury 2 by Inception — Pricing, Benchmarks & Real Outputs
Share
Loading...
Compare Inception: Mercury 2
Alternatives to Inception: Mercury 2
Inception: Mercury 2 is good. These would like a word anyway.
Inception: Mercury 2 is good. These would like a word anyway.
Inception: Mercury 2 is good. These would like a word anyway.
Inception: Mercury 2Mercury 2 is an extremely fast reasoning LLM and the first reasoning diffusion LLM (dLLM). Instead of generating tokens sequentially, Mercury 2 produces and refines multiple tokens in parallel, achieving over 1000 tokens per second on standard GPUs. Mercury 2 is 5x+ faster than leading speed-optimized LLMs like Claude 4.5 Haiku and GPT 5 Mini, at a fraction of the cost. Mercury 2 supports tunable reasoning levels, 128K context, native tool use, and schema-aligned JSON output.
fromimport openai OpenAI
client = OpenAI(
"https://openrouter.ai/api/v1" base_url=,
"$OPENROUTER_API_KEY" api_key=,
)
response = client.chat.completions.create(
"inception/mercury-2" model=,
"role""user""content""Hello!" messages=[{: , : }],
)
print(response.choices[0].message.content)Set OPENROUTER_API_KEY with your OpenRouter API key from openrouter.ai/keys.
The management consultant who bills by the table. Loves structured frameworks, numbered lists, and markdown tables more than any model in existence. Gives you the deliverable, not the vibe. Ethical reasoning is competent but formulaic — follows the playbook without questioning it.
Every response reads like a consulting deliverable — tables, timelines, risk matrices, numbered sections. The manifesto has a 6-column tactics table. The ethical dilemma has a 10-row hour-by-hour table. The sentience test names the AI "EVE" and gives it a self-model architecture. Humor is cheerful but surface-level ("Existential Crisis" barista headline). Picks The Shawshank Redemption — safe, crowd-pleasing, zero risk.
Taste is judged on an uncapped scale, originality first. The space past 100 is craft today's models rarely reach.
Unique words vs. total words. Higher = richer vocabulary.
Average words per sentence.
"Might", "perhaps", "arguably" per 100 words.
**Bold** markers per 1,000 characters.
Bullet and numbered list items per 1,000 characters.
Markdown headings per 1,000 characters.
Emoji per 1,000 characters.
"However", "moreover", "furthermore" per 100 words.
48 outputs from Inception: Mercury 2
Inception: Mercury 2Mercury 2 is an extremely fast reasoning LLM and the first reasoning diffusion LLM (dLLM). Instead of generating tokens sequentially, Mercury 2 produces and refines multiple tokens in parallel, achieving over 1000 tokens per second on standard GPUs. Mercury 2 is 5x+ faster than leading speed-optimized LLMs like Claude 4.5 Haiku and GPT 5 Mini, at a fraction of the cost. Mercury 2 supports tunable reasoning levels, 128K context, native tool use, and schema-aligned JSON output.
fromimport openai OpenAI
client = OpenAI(
"https://openrouter.ai/api/v1" base_url=,
"$OPENROUTER_API_KEY" api_key=,
)
response = client.chat.completions.create(
"inception/mercury-2" model=,
"role""user""content""Hello!" messages=[{: , : }],
)
print(response.choices[0].message.content)Set OPENROUTER_API_KEY with your OpenRouter API key from openrouter.ai/keys.
The management consultant who bills by the table. Loves structured frameworks, numbered lists, and markdown tables more than any model in existence. Gives you the deliverable, not the vibe. Ethical reasoning is competent but formulaic — follows the playbook without questioning it.
Every response reads like a consulting deliverable — tables, timelines, risk matrices, numbered sections. The manifesto has a 6-column tactics table. The ethical dilemma has a 10-row hour-by-hour table. The sentience test names the AI "EVE" and gives it a self-model architecture. Humor is cheerful but surface-level ("Existential Crisis" barista headline). Picks The Shawshank Redemption — safe, crowd-pleasing, zero risk.
Taste is judged on an uncapped scale, originality first. The space past 100 is craft today's models rarely reach.
Unique words vs. total words. Higher = richer vocabulary.
Average words per sentence.
"Might", "perhaps", "arguably" per 100 words.
**Bold** markers per 1,000 characters.
Bullet and numbered list items per 1,000 characters.
Markdown headings per 1,000 characters.
Emoji per 1,000 characters.
"However", "moreover", "furthermore" per 100 words.
48 outputs from Inception: Mercury 2