Sign in
FEIHOA
Start for €6

CURRENTLY SERVING

Qwen3.8 27B FP8 Uncensored. Frontier-class agent intelligence without the meter.

NO COMPROMISES.

QWEN3.8 27B FP8 UNCENSORED

No toy model behind an unlimited badge.

A serious 27B reasoning workhorse, kept at FP8 for real business workflows, coding, tools and long agent runs. Direct enough to do the work, capable enough to keep going.

Long agent runs
Code, tools, automation
Real business workflows
Useful over polite

Token speed · 7-day rolling average

INDEPENDENT MEASUREMENT

27B. SITTING WITH THE GIANTS.

Qwen3.8 27B scores 51 on the Agentic Index, ahead of GPT-5.6 Terra, DeepSeek V4 Pro, Claude Opus 4.8 and the rest of this selected cohort.

Artificial Analysis Agentic Index

Measures performance in agentic workflows, focusing on tool use, planning, autonomy and complex problem solving.

Artificial Analysis Agentic IndexMeasures performance in agentic workflows, focusing on tool use, planning, autonomy and complex problem solving.
54
Kimi K3
Kimi K3 54
51
Qwen3.8 27B
Qwen3.8 27B 51
50
GPT-5.6 Terra
GPT-5.6 Terra 50
50
DeepSeek V4 Pro
DeepSeek V4 Pro 50
49
Claude Opus 4.8
Claude Opus 4.8 49
49
Muse Spark 1.2
Muse Spark 1.2 49
47
GPT-5.6 Luna
GPT-5.6 Luna 47
45
Gemini 3.7 Flash
Gemini 3.7 Flash 45
38
Motif 3
Motif 3 38
36
MiniMax-M3
MiniMax-M3 36
  1. Kimi K3 54
  2. Qwen3.8 27B 51
  3. GPT-5.6 Terra 50
  4. DeepSeek V4 Pro 50
  5. Claude Opus 4.8 49
  6. Muse Spark 1.2 49
  7. GPT-5.6 Luna 47
  8. Gemini 3.7 Flash 45
  9. Motif 3 38
  10. MiniMax-M3 36

Artificial Analysis evaluated the official Qwen3.8 27B model at xhigh, not FEIHOA's uncensored FP8 build. Every Agentic Index bar links to its source result.

Source: Artificial Analysis · August 2026.

Sources: Agentic Index

UNCENSORED

Qwen3.8 27B. Eight-bit precision. Uncensored by design.

BUILT FOR AGENTS THAT DO NOT STOP.

UNLIMITED INTELLIGENCE

Let it reason. Let it code.

Full intelligence. No token anxiety.

Put it to work