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.
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
No measurements yet
We haven't recorded enough API calls to show availability stats for this model. Data appears once the model starts receiving live traffic.
Tokonomix benchmark verdicts
CogView-4 maintains steady image generation performance in second window
CogView-4 continues to demonstrate consistent image generation capabilities in its second benchmark window, showing no significant performance changes from its debut. The model maintains its position as a competent image generation solution with stable output quality. Users can expect reliable performance for standard image generation tasks, though the lack of measurable improvement suggests the model has settled into a steady state rather than showing rapid development. The absence of any performance degradation is noteworthy, indicating solid engineering and consistent model behavior across evaluation periods. For teams evaluating image generation options, CogView-4 presents as a stable choice that delivers predictable results without surprising regressions. The model's continued availability and unchanged capabilities make it suitable for production workflows where consistency is valued. However, users seeking cutting-edge improvements or new features will find this window uneventful. The benchmark data reveals neither breakthrough advances nor concerning declines, positioning CogView-4 as a reliable if unremarkable option in the current image generation landscape.
Quality
—
Latency p50
—
Test runs
0