Meta-Llama-3_3-70B-Instruct
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
$0.6700
input / 1M
— stable
$0.6700
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
100.0%
n=4
Last 30 days
100.0%
n=85
Median response time
119,154ms
n=85
Based on 475 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)
85
OK responses (30d)
85
Total calls (7d)
4
OK responses (7d)
4
Tokonomix benchmark verdicts
Quality recovers to 97.7 as latency improves 27%, coding performance dips
Meta-Llama-3.3-70B-Instruct demonstrates significant recovery in this benchmark window, climbing from 95.3 to 97.7 in overall quality while simultaneously improving latency performance by 27 percent, dropping from 10.5 seconds to 7.7 seconds at the median. This represents a notable operational improvement, bringing response times closer to competitive levels. However, the quality score remains below the 99.2 peak observed two windows ago, indicating the model has not fully returned to its previous performance ceiling. Category-level analysis reveals a mixed picture: multilingual capabilities have reached perfect scores at 100, surpassing the previous 98, while creative tasks score 95. Most notably, coding performance has declined from a perfect 100 to 98, representing the first observed weakness in this traditionally strong category. The reasoning category was not measured in the current window, making direct comparison unavailable. Users should expect reliable multilingual performance and improved response times, though those requiring maximum coding accuracy may notice subtle degradation from peak performance. The simultaneous improvement in speed and partial quality recovery suggests infrastructure optimizations, though the model has not yet achieved its demonstrated quality potential.
Quality
97.7
Latency p50
7,683 ms
Test runs
5