Qw
qwen3-30b-a3b-fp8 Benchmark & Insights
Qwen Cloudflare Workers AI
Updated Jul 18, 2026 All models
Sample size
156 runs
in window
Accuracy
49.4%
consensus match · 160d
Confidence
92%
over 161 runs
Window end
Jul 18, 2026
most recent run
Input price
$0.05/MTok
prompt tokens
Output price
$0.34/MTok
completion tokens
Model insights
- 01 Below a coin flip against consensus and stuck on "medium" (its most common answer, 85 of 155 days) regardless of conditions, with 103 overratings vs 2 under and ~40 false "unsafe" days.
- 02 The divergence log reads like a metronome — multiple misses every single week from February through July.
Recent forecasts