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
—
Last 30 days
100.0%
n=4
Median response time
6,509ms
n=4
Based on 361 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)
4
OK responses (30d)
4
Total calls (7d)
0
OK responses (7d)
0
Tokonomix benchmark verdicts
Grok 4.20 maintains consistent performance across benchmark windows
Grok 4.20 Non-Reasoning demonstrates stable performance between benchmark windows, with no significant changes in capability scores. The model continues to offer JSON output, tool use, and vision capabilities as previously established. Across coding benchmarks, the model maintains identical scores: 70.3% on HumanEval, 83.9% on MBPP, and 48.8% on SWE-bench Verified. Mathematical reasoning remains steady at 83.5% on MATH-500 and 89.5% on GSM8K. Instruction following shows consistent performance at 81.0% on IFEval with 87.1% on the strict variant. The model's multimodal capabilities show unchanged scores, with 59.8% on MMMU and 76.1% on MathVista. Graduate-level reasoning tasks maintain previous performance levels at 57.1% on GPQA Diamond. This benchmark window reflects operational stability rather than advancement. Users can expect consistent output quality matching the previous window, with established strengths in mathematical problem solving and instruction adherence. The model continues to serve as a reliable non-reasoning variant within the Grok family, maintaining its position without notable gains or regressions.
Quality
—
Latency p50
—
Test runs
0
Grok 4.20 Non-Reasoning (0309)
by xAI
- Context window
- 1.048576M tokens
- Input price
- $1.87 / 1M
- Output price
- $3.74 / 1M
- Tier
- Tier B — Production
- Modality
- Text + vision
- API type
- REST · streaming
- Benchmark runs
- 209
More from xAI