Mercury 2.5 Preview is Inception's latest and most intelligent diffusion language model. Instead of generating tokens strictly one at a time, it produces and refines multiple tokens in parallel, reaching up to 1,107 tokens per second on standard GPUs. It delivers a 10+ point intelligence gain over Mercury 2, with tunable reasoning, parallel tool calls, schema-aligned JSON output, and a 260K context window. It is built for latency-sensitive production work such as search agents, voice pipelines, customer support, rapid coding iteration, and coding subagents.
Added Sep 1, 2026
Context Window
260.0K
Max Output
65.5K
Input Price (Auto)
$0.040/1M
Output Price (Auto)
$0.15/1M
Cache Read (Auto)
$0.0040/1M
Capabilities
Benchmarks
Benchmarks
Performance metrics and benchmarks
No benchmark data is available yet for this model.
Providers
Auto routing is available for this model. Explicit provider selection is not available.
Loading provider options…
Related text models
Compare Mercury 2.5 Preview with similar models from the same provider or model family.
Mercury 2
mercury-2Inception Labs' fastest reasoning model with tool calling and structured outputs support.
Mercury Coder Small
mercury-coder-smallModel by Inception AI. A diffusion large language model that runs incredibly quickly (500+ tokens/second) while matching Claude 3.5 Haiku and GPT-4o-mini. 1st in speed on Copilot arena, and matching 2nd in quality.
Abliterated Model Large V2
abliteration-ai/abliterated-model-large-v2Abliteration.ai's default unrestricted large text reasoning model is derived from GLM-5.3 and weight-modified to reduce refusals compared with the base model. It is intended for harder reasoning and evaluation workloads, with automatic prompt caching and a one-million-token context window.
Granite 4.2 8B
ibm-granite/granite-4.2-8bIBM Granite 4.2 8B is an Apache 2.0-licensed dense model with native step-by-step reasoning and specialized training for agentic work. It can plan before acting, sequence tools, navigate codebases, work in terminals, and verify results across coding, search, mathematics, science, and complex instruction-following tasks.
GLM 5.3 TEE
TEE/glm-5.3GLM-5.3 is Z.AI's open-weight reasoning model for complex software engineering, autonomous agents, vulnerability research, and long-horizon tasks. This text-only deployment runs inside a Phala Trusted Execution Environment with Redpill attestation and signed completion receipts.
GLM 5.3 Flash TEE
TEE/glm-5.3-flashGLM-5.3 Flash is Z.AI's natively multimodal 320B MoE reasoning model with 18B active parameters, served by Phala inside a Trusted Execution Environment with Redpill attestation and signed completion receipts.