MetaLlama 3.3 Instruct 70B
- Released
- Dec 6, 2024
- Tracked since
- 2026-07-23
Intelligence index7.7Ranks 462 of 664 models with this metric
Coding index11.9Ranks 229 of 259 models with this metric
Agentic index—Not tested
Capability profile
5 axes vs. site averageLlama 3.3 Instruct 70B
Site average
Some dimensions have no data yet — marked on the chart
Pricing
per 1M tokens- Input
- $0.710
- Output
- $0.720
- Cache read
- $0.710
Performance
- Output throughput
- 85 tokens/sec
- Time to first token
- 1.68s
- First answer token
- 1.68s
- End-to-end response
- 7.55s
History
9Intelligence index
9.4→7.7
Output
$0.710→$0.720
Output throughput
80→85
Cheaper at this level
Models with a comparable intelligence index and a lower output price- Gemma 4 E4B (Reasoning)
GoogleIntelligence 8.9$0.100−86% - Gemma 4 E4B (Non-reasoning)GoogleIntelligence 7.5$0.100−86%
- Granite 4.2 3B
IBMIntelligence 9.1$0.120−83% - Qwen3.5 4B (Reasoning)
AlibabaIntelligence 13.1$0.150−79% - Qwen3.5 4B (Non-reasoning)AlibabaIntelligence 10.8$0.150−79%