Inception
Inception ranks #39 of 91 providers on BenchLeader by its best model, with 2 ranked models. Its best model is Mercury 2 at 46.9 on the BenchLeader Index, strongest in instruction following. Prices run from $0.375 to $0.375 per million tokens blended; Mercury 2 is the cheapest ranked option. The fastest is Mercury 2.5 at 689 tokens per second. It has shipped 5 dated releases, roughly one every 38 days; the last was 19 Oct 2026, so the next would be due between 10 Nov 2026 and 11 Dec 2026 on that pattern.
Diffusion-LLM pioneer; Mercury models generate tokens in parallel for very high output speed.
- Provider rank
- #39
- by best model
- Best index
- 46.9
- Mercury 2
- Ranked models
- 2
- 0 open weights
- Cheapest ranked
- $0.375/M
- Mercury 2
Release timeline
Each model’s best configuration by release date. The top eight are labelled; hover any point for its name.
- Inception2
Next release
Over the last two years Inception shipped 5 distinct models, one every 38 days on median. The latest arrived on 19 Oct 2026.
If the cadence holds, the next one lands between 10 Nov 2026 and 11 Dec 2026.
An estimate from past release dates only. It knows nothing about roadmaps or announcements.
Fastest ranked model: Mercury 2.5 at 689 tok/s.
Models
Every Inception model with a result, one row per model; switch to all configurations to see each reasoning effort.
| # | ||||||||
|---|---|---|---|---|---|---|---|---|
| 1 | Mercury 2 | 46.9±6 | $0.375 | 451 | 7.79 s | 128k | 2026-02-20 | 2026-10-23 |
| 2 | Mercury 2.5 | 45.3±7 | $0.375 | 689 | 3.71 s | 260k | 2026-09-08 | 2026-09-08 |
| 3 | Mercury Coder | – | – | – | – | – | – | – |
| 4 | NewMercury 3high | – | $0.163 | 267 | 0.92 s | 1M | 2026-10-19 | – |
| 5 | Mercury Edit 2 | – | $0.375 | – | – | 32k | 2026-03-30 | – |
| 6 | Mercury Voice | – | $0.675 | – | 0.32 s | 128k | 2026-09-29 | – |
Latest news
- Inception says Mercury 3 beats GPT-6 Luna and Claude Haiku 5.5 at up to 8x the speedSiliconANGLE · 20 Oct 2026
- Inception Releases Mercury 3, Its Flagship Diffusion LLMUnite.AI · 19 Oct 2026
- RhymeQ Joins NVIDIA Inception Program and Announces Technical Collaboration on NOOA Agentic AI FrameworkQuantum Computing Report · 23 Oct 2026
- Inception's Mercury 3 diffusion model claims to beat GPT-6 Luna and Claude Haiku 5.5 at a fraction of the waitThe Decoder · 21 Oct 2026
- Inception Launches Mercury 3, a Diffusion AI Model Generating Over 1,000 Tokens per SecondentARABI · 20 Oct 2026
- Inception Launches Mercury 3 Diffusion AI Model With 1,068 Tokens Per SecondKonsulteer · 21 Oct 2026
- Inception’s Mercury 3: 40x Bigger, 8x Faster [2026]shattered.io · 20 Oct 2026
- Inception’s Mercury 3 Release Challenges Sequential AI Generationremio · 21 Oct 2026