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.
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=9
Last 30 days
100.0%
n=17
Median response time
3,987ms
n=17
Based on 137 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)
17
OK responses (30d)
17
Total calls (7d)
9
OK responses (7d)
9
Tokonomix benchmark verdicts
Grok 4.20 Non-Reasoning debuts with strong multimodal performance
Grok 4.20 Non-Reasoning enters the benchmark landscape as xAI's latest offering, demonstrating competitive performance across key evaluation metrics. The model achieves a 78.8% score on MMLU, positioning it solidly in the upper tier of current language models, though trailing the leading systems that exceed 85%. On mathematical reasoning tasks via MATH, it scores 71.9%, showing respectable capability but remaining below the frontier models approaching or surpassing 90%. The model delivers 85.5% on MMLU-Pro and 74.9% on GPQA Diamond, indicating strong performance on more challenging academic benchmarks. Coding ability registers at 74.2% on HumanEval, a competent showing for general programming tasks. Multimodal capabilities are notable, with 78.2% on MMMU demonstrating solid vision-language understanding. The model supports JSON structured outputs and tool calling functionality, making it practical for application development. While this baseline verdict cannot assess performance changes, Grok 4.20 Non-Reasoning establishes itself as a capable general-purpose model with particular strength in multimodal tasks, suitable for production use cases requiring reliable but not necessarily state-of-the-art performance across diverse domains.
Quality
—
Latency p50
—
Test runs
0