{"schema":"postcutoff/event@1","as_of":"2026-10-20T07:28:00+02:00","url":"https://postcutoff.com/e/2026-10-19-inception-mercury-3/","md":"https://postcutoff.com/e/2026-10-19-inception-mercury-3/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-10-19-inception-mercury-3","date":"2026-10-19","date_precision":"day","short_title":"Inception releases Mercury 3, a diffusion LLM with a 1M-token context that it says beats GPT-6 Luna and Claude Haiku 5.5 at 1,000+ tokens/s","deck":null,"takeaway":"On Oct 19, 2026 Inception released Mercury 3, its flagship diffusion language model.","category":"model-release","category_label":"Model releases","importance":3,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":[{"n":1,"title":"Inception: Diffusion at Scale: Introducing Mercury 3","url":"https://www.inceptionlabs.ai/blog/introducing-mercury-3","type":"official","group":"primary","domain":"inceptionlabs.ai"},{"n":2,"title":"Inception docs: Models and pricing (mercury-3)","url":"https://docs.inceptionlabs.ai/get-started/models","type":"docs","group":"primary","domain":"docs.inceptionlabs.ai"},{"n":3,"title":"The Deep View: Why Inception's Mercury 3 upends AI economics","url":"https://www.thedeepview.com/articles/why-inception-s-mercury-3-upends-ai-economics","type":"press","group":"press","domain":"thedeepview.com"}],"official":2,"filed":"2026-10-20","updated":"2026-10-20","orgs":["Inception"],"title":"Inception releases Mercury 3, a diffusion LLM with a 1M-token context that it says beats GPT-6 Luna and Claude Haiku 5.5 at 1,000+ tokens/s","summary":"On Oct 19, 2026 Inception released Mercury 3, its flagship diffusion language model. It has a 1M-token context and runs at 1,068 tokens/s on Nvidia GPUs. Inception says it outperforms GPT-6 Luna and Claude Haiku 5.5 while running 4–8x faster, but publishes no benchmark scores. Launch price is $0.10 / $0.35 per 1M input/output tokens (50% off the $0.20 / $0.70 list price).","key_facts":["API id: mercury-3 (Inception API, OpenAI-compatible; also OpenRouter as inception/mercury-3, Baseten and Vercel)","Context 1M tokens (Mercury 2.5: 260K); max output 131,072 tokens; reasoning effort instant/low/medium/high (Inception docs)","Speed: 1,068 tokens/s on Nvidia GPUs, per Inception; claimed 4–8x faster than GPT-6 Luna, Claude Haiku 5.5 and GLM-5.3 Flash","Price: $0.20 input / $0.70 output per 1M tokens, 50% off at launch ($0.10 / $0.35); cached input 90% off. The Deep View: launch input price equals Haiku 5.5 and GPT-6 Luna, output 35% cheaper","Claims without published scores: beats GPT-6 Luna and Claude Haiku 5.5 'across a range of benchmarks'; completed 50% of 50 live Mind2Web browser tasks, 'significantly more than GPT-6.1-Sol' (Sol's score not given)","Scale: over 40x the total parameters of Mercury 1 (Feb 2025); count not disclosed","The Deep View: Inception says swarms of Mercury 3 agents were used for recursive self-improvement that raised speed a further 16%; Inception's own blog post does not repeat this claim","CEO Stefano Ermon (The Deep View): 'Mercury 3 is the most capable diffusion-based language model on the market'; he said Inception is working on making the launch price permanent"],"key_numbers":[],"tags":["llm","diffusion-llm","fast-inference","api","agents"],"science":null,"body_md":"## What happened\n\nInception, the Stanford-founded startup behind the first commercial diffusion LLMs, released Mercury 3 on Oct 19, 2026. Instead of\nwriting one token at a time, the model drafts and refines many tokens in parallel, which Inception says gives it over 1,000 tokens\nper second. The release adds a 1M-token context window and four reasoning-effort levels, and is sold at half price at launch.\n\nInception's comparisons are claims: the blog post gives no benchmark table, and the only task figure (50% of 50 live Mind2Web\ntasks) has no matching number for the GPT-6.1 Sol model it is compared with. No independent measurement was available on\nOct 20.\n\n## Why it matters\n\nDiffusion LLMs are the main alternative to autoregressive decoding. If Mercury 3's claims hold, a small-model tier of speed and\nprice now comes with long context and agentic ability close to the cheap tiers of OpenAI and Anthropic.","disputed":[],"related":[],"people":[{"id":"stefano-ermon","name":"Stefano Ermon","url":"https://postcutoff.com/person/stefano-ermon/"}],"posts":[],"videos":[],"models":[{"id":"mercury-3","name":"Mercury 3","url":"https://postcutoff.com/m/mercury-3/"}],"changes":[{"date":"2026-10-20","type":"filed","text":"Created (Techmeme / The Deep View; Inception blog and docs read)"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-10-20","run":null,"sources_read":"Techmeme / The Deep View; Inception blog and docs read","updated":"2026-10-20","human_review":null,"version":null},"gaps":[{"model_id":"gpt-6-astra","name":"GPT-6 Astra","cutoff":"2026-04","days_after":172,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":111,"in_training_data":false},{"model_id":"gemini-4-argon","name":"Gemini 4 Argon","cutoff":"2026-03","days_after":202,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":141,"in_training_data":false}],"short_url":null}