Qwen3-32B
Speed analysis
Latency measured across all benchmark runs. P50 (median) and P95 (95th percentile) give a realistic picture of response speed under normal and peak load.
Quality scores
Evaluation results from judge-model scoring across diverse task categories. Scores reflect coherence, accuracy and instruction-following.
Pricing history
Direct provider rates per million tokens, plus a typical-conversation cost estimate.
Pricing over time
Input & output per 1M tokens · step-line = price changes
$0.0800
input / 1M
— stable
$0.2300
output / 1M
— stable
Tokens per second
Throughput in tokens per second, derived from measured P50 latency. Higher is better; fluctuations track provider-side load.
Estimated from P50 latency × 200 output tokens — the absolute number depends on this assumption; the trend is what matters.
Capabilities
Availability
Availability
How often this model answers when we call it — measured across real API requests and live tests over the last 30 days. This is separate from quality: these numbers only tell you whether the model responds, not how good the answer is.
Last 7 days
—
Last 30 days
100.0%
n=33
Median response time
145,961ms
n=33
Based on 413 measurements over the last 30 days.
Technical details
Only live API calls and live-test requests count — internal probes and benchmark runs are excluded.
Calls with a custom API key (BYOK) are excluded: those failures are key-specific, not a sign of model downtime.
Failed calls are NOT included in quality scores — quality is measured on successful responses only. Availability and quality are independent signals.
Median response time (p50) across successful calls with a recorded duration. Outliers (very slow or very fast calls) pull the median less than the average.
Total calls (30d)
33
OK responses (30d)
33
Total calls (7d)
0
OK responses (7d)
0
Tokonomix benchmark verdicts
Qwen3-32B quality drops to 73.4 with sharp creative performance decline
Qwen3-32B has experienced a significant performance regression in this benchmark window, with overall quality falling from 96.6 to 73.4 points. The most dramatic decline occurred in creative tasks, which scored just 40 points compared to previous strong performance levels. This represents a substantial capability loss in content generation and creative reasoning tasks. Coding performance remains the model's strongest area at 94 points, showing resilience despite the overall quality drop. Multilingual capabilities scored 86, indicating reasonable but diminished performance compared to the previous window's 97. The absence of a reasoning score in current results, compared to 95 previously, suggests potential issues with test completion or capability regression in logical tasks. Latency has also degraded noticeably, with p50 response times increasing 43 percent from 17167ms to 24595ms. Users can expect waits of approximately 25 seconds for typical requests. The combination of reduced quality and slower response times indicates potential infrastructure issues or model configuration changes that have impacted both performance dimensions. Teams relying on creative output or time-sensitive applications should evaluate whether current performance meets their requirements.
Quality
73.4
Latency p50
24,595 ms
Test runs
5