Mercury 2.5
NewStandardInception's latest diffusion LLM and the fastest reasoning model available, producing and refining tokens in parallel instead of one at a time.
Over 1,100 tokens per second on standard GPUs with a 10+ point intelligence jump over Mercury 2, comparable to cost-optimized frontier models. 256K-token context with tunable reasoning, parallel tool calls, and schema-aligned JSON output. Built for production work where latency compounds: search agents, voice pipelines, and coding subagents.
▼
Mercury 2.5 is somewhat expensive at $0.04/M tokens input and $0.15/M tokens output. Cached input reads are priced at $0.004/M tokens, making repeated context cheaper. The model supports extended reasoning, a 260K context window.
Benchmarks
Specifications
Performance
Providers
Misc
Compare with
GPT 6 Astravs Mercury 2.5GPT 5.6 Lunavs Mercury 2.5
Claude Fable 5.1vs Mercury 2.5Claude Opus 5vs Mercury 2.5Claude Sonnet 5vs Mercury 2.5Claude Haiku 4.5vs Mercury 2.5
Gemini 3.8 Flashvs Mercury 2.5Gemini 3.5 Flashvs Mercury 2.5Gemini 3.5 Flash-Litevs Mercury 2.5Gemma 4 31Bvs Mercury 2.5
Grok 4.6vs Mercury 2.5Grok Build 0.1vs Mercury 2.5Grok 4.1 Fastvs Mercury 2.5
Muse Spark 1.3vs Mercury 2.5
DeepSeek V4 Pro (0813)vs Mercury 2.5DeepSeek V4 Flashvs Mercury 2.5
MiniMax M3vs Mercury 2.5
GLM 5.3vs Mercury 2.5GLM 5.3 Flashvs Mercury 2.5
Qwen3.8 Max (0902)vs Mercury 2.5Qwen3.7 Plusvs Mercury 2.5Qwen3 Coder Nextvs Mercury 2.5
Nemotron 3 Ultravs Mercury 2.5Frequently Asked Questions about Mercury 2.5▼
When was Mercury 2.5 released?
▼
Mercury 2.5 was released on September 8, 2026.
Who created Mercury 2.5?
▼
Mercury 2.5 was created by Inception AI.
How much does Mercury 2.5 cost?
▼
Mercury 2.5 costs $0.04/M input tokens and $0.15/M output tokens.
What is Mercury 2.5 API pricing?
▼
The Mercury 2.5 API is priced at $0.04 per million input tokens and $0.15 per million output tokens. You can access Mercury 2.5 through idapt.app alongside 200+ other AI models in one workspace.
Is Mercury 2.5 a reasoning model?
▼
Yes, Mercury 2.5 is a reasoning model. It can think step-by-step through complex problems before providing an answer, often yielding better results on difficult tasks like math, coding, and multi-step logic.
Does Mercury 2.5 support image or vision input?
▼
No, Mercury 2.5 does not currently support image or vision input. It is a text-only model.
Does Mercury 2.5 support audio?
▼
No, Mercury 2.5 does not support audio input.
What is the context window of Mercury 2.5?
▼
Mercury 2.5 has a context window of 260K tokens, meaning it can process approximately 195,000 words of context at once.
Is Mercury 2.5 open source?
▼
Mercury 2.5 is a proprietary model developed by Inception AI. The model weights and training data are not publicly available.
Where can I use Mercury 2.5?
▼
You can chat with Mercury 2.5 on idapt.app — no separate API key required. idapt is an AI workspace with 200+ models, agents, computers, and tasks. You can also access Mercury 2.5 through Inception AI's own API or platform.