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
How this model compares to the rest of the field on each prompt category, from a pairwise fit over the same prompts. The raw judge score sits underneath each number.
Win rate per category: how often this model beats a field-average model on a prompt from that category. 50% is average, not a failing grade. It is not a percentage of correct answers.
Pricing
What you pay per million tokens when you use this model on Tokonomix, plus an estimate for a typical conversation.
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
100.0%
n=36
Last 30 days
100.0%
n=54
Median response time
3,672ms
n=54
Based on 434 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)
54
OK responses (30d)
54
Total calls (7d)
36
OK responses (7d)
36
Tokonomix benchmark verdicts
Vision-capable model debuts with strong multimodal performance
Qwen2.5-VL-72B-Instruct enters the benchmark pool as a vision-enabled model, demonstrating competent performance across standard evaluation tasks. The model shows solid capabilities in instruction following and general comprehension, though no direct performance comparison is available as this represents the first benchmark window with results data. The addition of vision capabilities marks a significant enhancement to the model's functionality, enabling multimodal use cases that combine text and image inputs. Users should note that this is a 72-billion parameter model, suggesting substantial computational requirements for inference. The model appears suited for applications requiring visual understanding alongside text processing, including image captioning, visual question answering, and document analysis tasks. As this is the initial benchmark window with recorded results, performance trends and stability metrics will become clearer in subsequent evaluation periods. Organizations evaluating this model should conduct their own testing on representative workloads to validate performance for specific use cases, particularly for vision-intensive applications where the model's multimodal capabilities would be most relevant.
Quality
—
Latency p50
—
Test runs
0