Skip to content
NewA calmer, faster ZeroHour
StoryModel releaseinfo

Inception releases Mercury 3 diffusion model at 1,068 tokens/sec

2 sources · 2 articlesFirst reported Updated ()1
What's new

First merged summary.

Merged summary · x-ai/grok-4.7 · rewritten as coverage arrivesSummaries by Vani AI

Inception has released Mercury 3, a diffusion model it says runs at 1,068 tokens per second, without independent verification.

Inception announced Mercury 3, a diffusion language model it calls the most capable of its kind released so far. The company says its tests give Mercury 3 eight of 12 wins against GPT-6 Luna and Claude Haiku 5.5, including 76.84 on IFBench and 51.08 on SWE-Bench Atlas, at 1,068 tokens per second versus about 240 and 130. A shorter overview says quality is on par or better, and Claude Haiku 5.5 won DeepSWE and long-text retrieval tests, according to The Decoder. It is pitched for agents, search, support and tutoring, with launch pricing cut in half, regular rates of $0.20 per million input tokens and $0.70 output, and availability on Inception, OpenRouter, Baseten and Vercel. MarkTechPost's captured page was only a cookie notice, so it does not confirm architecture or terms. No independent verification is reported.

Key facts
  • Company tests: eight of 12 wins versus GPT-6 Luna and Claude Haiku 5.5.
  • IFBench score 76.84, 15 points above GPT-6 Luna, according to Inception.
  • SWE-Bench Atlas 51.08 versus 33.33 and 32.26 for the rivals.
  • Claude Haiku 5.5 won DeepSWE and long-text retrieval tests.
  • Regular prices: $0.20 per million input tokens and $0.70 output.
  • Offered on Inception, OpenRouter, Baseten and Vercel; launch prices halved.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

What each outlet adds

beyond the shared facts
  • The DecoderCompany benchmarks, mixed results, pricing, use cases and availability.
  • MarkTechPostCaptured page was only a cookie notice, confirming no extra terms.
Vendors
Inception
Organizations
OpenRouterBasetenVercelInception
AI models
Mercury 3GPT-6 LunaClaude Haiku 5.5

Coverage

oldest first · one row per article · UTC
  1. · 5d ago MarkTechPostFirst report

    Inception releases Mercury 3 diffusion LLM at 1,068 tokens per second

    The page body is a cookie banner, so only the stated generation speed is available, not architecture or availability.

  2. · 4d ago The Decoder

    Inception claims Mercury 3 beats GPT-6 Luna and Claude Haiku 5.5

    Inception's own tests give Mercury 3 eight of 12 wins and 1,068 tokens per second, with no independent verification yet.