Mercury 2.5
Inception · Released Sep 2026
Conditional. Mercury 2.5 is Inception's diffusion LLM, and raw output speed is the reason to pick it: Artificial Analysis ranks it among the fastest models it measures, and it ships tunable reasoning, parallel tool calls and structured outputs. The deciding caveat is quality: Artificial Analysis places its intelligence just below the median, and Inception's parity claims name no benchmark. Suits latency-bound voice agents, RAG pipelines and coding subagents, not hard reasoning.
Is it right for you?
Good for
- Latency-bound work: Artificial Analysis measures 636.5 output tokens per second, #3 of 175 models; Inception cites 1,107 tokens per second on NVIDIA GPUs
- Agent subtasks needing tool use: parallel tool calls, schema-aligned JSON and a reasoning_effort parameter
- Search agents, RAG pipelines, voice agents and coding subagents, the use cases Inception names
- Long inputs up to 260K tokens with up to 65,536 output tokens
Not good for
- Hard reasoning where quality beats speed: Intelligence Index 12, which Artificial Analysis labels below average among reasoning models in a similar price tier (median 12)
- Image, audio or video input; text only
- Time-to-first-token-critical calls: AA measures 3.87 s TTFT vs a 2.20 s median
How it performs by task
High-throughput generation
636.5 tok/s measured by Artificial Analysis, #3 of 175
Tool-calling subagents
Parallel tool calls and structured outputs are documented; no agentic benchmark published
Complex reasoning
Intelligence Index 12, below average among similarly priced reasoning models (median 12) per Artificial Analysis
Pricing
Input
$0.20 / 1M tokens
Output
$0.75 / 1M tokens
Context
260K ctx; 80% off at launch: $0.04/$0.15
Benchmarks
No verdict changes yet
The clock starts day one. Changes land here as our verdict evolves.
Sources
Verification log
- Pricing· No changes
Automated agent
- Pricing· No changes
Automated agent
- Profile· Updated
Automated agent
Inception list pricing unchanged ($0.20/$0.75; launch $0.04/$0.15; 260K ctx). Artificial Analysis now shows 636.5 tok/s, #3 of 175 (was 843, #2 of 174), TTFT 3.87 s vs 2.20 s median (was 3.08 s vs 2.19 s), Intelligence Index 12 with median 12 (was median 13). Benchmarks, strengths, task ratings, sources refreshed; summary rewrite staged for publish; draft news created.
- Pricing· No changes
Automated agent