The Known Good Updated 29 Jul 2026
Meta ·Open weights (restricted use) ·Released December 2024·Tracked since 25 Jul 2026

Meta: Llama 3.3 70B Instruct — quality, speed & price analysis

Compare with another model Provider benchmarks Data sources
Quality
#53 of 62
34based on 4 of 7 evaluations
Known Good Index
Speed
#187 of 317
43.5
Output tokens per second
Price
#74 of 370
$0.20
Blended per 1M tokens
In $0.13 · Out $0.40
Cache hit price
unranked
—
per 1M tokens
Context
#279 of 375
131k
tokens (~200 pages)
Summary

Meta: Llama 3.3 70B Instruct is in the lowest quarter of the models we track for quality, and inexpensive for its standing, and notably slow. It accepts text input, outputs text, and supports a 131k-token context window.

It scores 34 on the Known Good Index, well below the tracked average of 64, from 4 of the seven component evaluations. Measured against the tracked average, its strongest result is LiveBench (50 against 52) and its weakest is Mock AIME 2024–25 (5 against 53).

Pricing is $0.13 per 1M input tokens (below the $2.10 average) and $0.40 per 1M output (below the $9.62 average). At 44 tokens per second it sits in the lower middle of the tracked range.

Generated from tracked data on 29 Jul 2026 · how these comparisons are computed →

Technical specifications
ReasoningNo
Input modalitytext
Output modalitytext
Context window131,072 tokens
Max output128,000 tokens
WeightsOpen weights · Open weights (restricted use) (commercial use restricted)
Knowledge cutoffDecember 2023
Tool callingYes
Providers12 hosts · compare →
Highlights — where Meta: Llama 3.3 70B Instruct sits
Quality

Known Good Index · Higher is better · Composite of seven published evaluations

98
96
95
94
93
93
93
91
90
90
88
88
86
83
83
80
80
78
78
77
75
75
74
74
74
73
72
70
68
68
34
gpt-5.5-pre-release
Gemini 3.1 Pro
Preview
Gemini 3.5 Flash
Claude Opus 4.7
Qwen3.7 Max
DeepSeek V4 Pro
Kimi K2.6
GLM 5.2
Qwen3.6 Max
Preview
GPT-5.4
Claude Opus 4.6
GLM 5.1
o3 Mini
Gemini 3 Flash
Preview
gemini-3-pro-preview
GPT-5.2
Claude Sonnet 4.6
GPT-5
Qwen3.6 Plus
GLM 5
R1
deepseek-r1
claude-3-7-sonnet-20250219
GPT-5.1
o3
gemini-2.0-pro-exp-02-05
claude-opus-4-5-20251101
o1
claude-opus-4-20250514
GLM 4.7
Llama 3.3 70B
Instructthis model #53/62
Speed

Output tokens per second · Higher is better · OpenRouter endpoint telemetry

2,504
1,909
1,434
566
370
342
280
181
166
161
154
148
145
144
143
136
135
135
134
132
131
130
130
126
126
126
124
121
108
107
44
Morph V3 Large
Morph V3 Fast
Relace Apply 3
Grok 4.20
Multi-Agent
gpt-oss-safeguard-20b
Nano Banana 2 Lite
(Gemini 3.1 Flash Lite Image)
Mercury 2
Nano Banana
(Gemini 2.5 Flash Image)
Grok 4.20
Nano Banana 2
(Gemini 3.1 Flash Image Preview)
Llama 3.2 3B
Instruct
Qwen3 Next 80B A3B
Thinking
gpt-oss-120b
Gemini 3.6 Flash
Laguna XS 2.1
Muse Spark 1.1
Trinity Large
Thinking
o3 Mini High
Qwen3 30B A3B
Thinking 2507
gpt-oss-20b
Nano Banana 2
(Gemini 3.1 Flash Image)
Gemini 3.1 Flash
Lite Preview
GPT-5.6 Luna Pro
Qwen3 VL 8B
Thinking
Gemini 3.5 Flash
Llama 3.2 1B
Instruct
Gemini 3.1 Flash
Lite
Grok Build 0.1
Claude Opus 4.8
(Fast)
Nemotron 3 Nano
30B A3B
Llama 3.3 70B
Instructthis model #187/317
Price

Blended USD per 1M tokens · Lower is better · Published pricing via OpenRouter

$0.01
$0.02
$0.04
$0.04
$0.04
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.08
$0.09
$0.09
$0.09
$0.10
$0.10
$0.10
$0.10
$0.11
$0.20
Ling-2.6-flash
Mistral Nemo
Granite 4.0 Micro
Llama 3 8B Lunaris
Nex-N2-Mini
Qwen2.5 7B
Instruct
gpt-oss-20b
Qwen3.7 Flash
Mistral Small 3
Llama 3.1 8B
Instruct
MythoMax 13B
Nova Micro 1.0
Granite 4.1 8B
Gemma 3 4B
Command R7B
(12-2024)
GPT-5 Nano (batch)
gpt-oss-120b
Llama 3.2 1B
Instruct
Gemma 3 12B
Laguna XS 2.1
Gemma 3n 4B
Qwen3 30B A3B
Instruct 2507
Nemotron 3 Nano
30B A3B
Gemini 2.5 Flash Lite
(batch)
Phi 4
Hy3 preview
Ministral 3 3B
2512
Reka Edge
Nova Lite 1.0
Qwen3.5-9B
Llama 3.3 70B
Instructthis model #74/370 · off scale
Jump to section

Quality

How Meta: Llama 3.3 70B Instruct compares against every tracked model

Known Good IndexCapability indices
Known Good Index

Composite of GPQA Diamond, Mock AIME 2024–25, MATH Level 5, SWE-bench Verified, LiveBench, Humanity's Last Exam and Terminal-Bench, normalised 0–100. Scores are ingested from Epoch AI, LiveBench; we do not run these evaluations.

⇩
98
96
95
94
93
93
93
91
90
90
88
88
86
34
💡 A lightbulb marks a model its provider publishes as a reasoning model
The Known Good

Quality Breakdown

Meta: Llama 3.3 70B Instruct on each evaluation feeding the index, against the tracked average

Evaluation scores

Bar = Meta: Llama 3.3 70B Instruct · grey marker = average across tracked models · scores ingested from Epoch AI, LiveBench

GPQA Diamond47 / 63
Mock AIME 2024–255 / 53
MATH Level 542 / 55
LiveBench50 / 52
The Known Good

Arena Elo

Human preference rating from blind pairwise votes

Text Arena

Elo from blind pairwise votes, ingested from Arena

1275
Elo · 95% CI ±3
#241
of 386 ranked models
55k
votes counted
1318
style-controlled (#230)

Ingested 29 Jul 2026.

The Known Good

Price & Cost

Published API pricing by token type

BlendedInput vs outputQuality per dollar
Blended Price

USD per 1M tokens at a 3:1 input:output blend. Live from OpenRouter, cross-checked against models.dev. Lower is better.

⇩
$0.01
$0.02
$0.04
$0.04
$0.04
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.20
The Known Good
We publish blended price rather than cost-per-task. Cost-per-task requires token counts from an evaluation suite we do not run. Blended price is computed from published rates and is fully reproducible — see the pricing methodology.

Context Window

Maximum input tokens accepted

Context windowMax output tokens
Context Window

Maximum input context from provider metadata via models.dev and OpenRouter. Higher is better.

⇩
2.00M
2.00M
2.00M
2.00M
2.00M
1.31M
1.05M
1.05M
1.05M
1.05M
1.05M
1.05M
131k
The Known Good

Speed

Median output throughput across hosting providers

Output speedBy provider
Output Speed

Median output tokens per second across all tracked providers. Derived from OpenRouter endpoint telemetry. Higher is better.

⇩
2,504
1,909
1,434
566
370
342
280
181
166
161
154
148
44
The Known Good

Latency

Time to first token, median across providers

Latency (TTFT)

Seconds to first streamed chunk, median across tracked providers. Derived from OpenRouter endpoint telemetry. Lower is better.

⇩
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.3
0.3
0.6
The Known Good

Providers

Hosts serving Meta: Llama 3.3 70B Instruct, with their own price and speed

12 providers
Full provider benchmarks →
ProviderInput $/MOutput $/M BlendedSpeedTTFTContext
Groq $0.59 $0.79 $0.64 174 0.3s 131k
Google · first party $0.72 $0.72 $0.72 72 0.3s 128k
Crusoe $0.25 $0.75 $0.38 68 0.3s 131k
SambaNova $0.45 $0.90 $0.56 65 0.7s 16k
CoreWeave $0.71 $0.71 $0.71 52 0.3s 128k
Parasail $0.22 $0.50 $0.29 36 0.6s 131k
Novita $0.14 $0.40 $0.20 31 0.6s 6k
Cloudflare $0.29 $2.25 $0.78 24 0.7s 24k
Nebius $0.13 $0.40 $0.20 24 0.6s 131k
Together $1.04 $1.04 $1.04 21 1.1s 131k
AkashML $0.13 $0.40 $0.20 20 0.7s 131k
DeepInfra $0.10 $0.32 $0.15 13 0.7s 131k

Per-host pricing and telemetry from OpenRouter. Speed is output tokens per second; TTFT is time to first token.

The Known Good

History

How price and quality have moved since we began tracking this model

Quality & price over time

Since 25 Jul 2026

Known Good Index 34 · based on 4 of 7 evaluations
Blended price $0.20 per 1M · observed 29 Jul 2026
Output speed 44 tok/s · observed 29 Jul 2026
Time to first token 0.6s
29 Jul 2026 · price change
Meta: Llama 3.3 70B Instruct price up 27%
Blended 3:1 price moved from $0.16 to $0.20 per 1M tokens (OpenRouter).
28 Jul 2026 · price change
Meta: Llama 3.3 70B Instruct price down 22%
Blended 3:1 price moved from $0.20 to $0.16 per 1M tokens (OpenRouter).
The Known Good

Comparisons

Head-to-head pages pairing Meta: Llama 3.3 70B Instruct with other tracked models

Head-to-head pages are generated for the top 50 models by Known Good Index, which Meta: Llama 3.3 70B Instruct is not currently in. Its nearest tracked models by index are below.

mistral-small-2503 claude-3-opus-20240229 mistral-small-2501 OpenAI: GPT-4o-mini mistral-large-2411 claude-3-5-haiku-20241022 OpenAI: GPT-4o Microsoft: Phi 4
Sources for this page
Pricing & providers — OpenRouter Metadata — models.dev Benchmarks — Epoch AI (CC-BY 4.0)Benchmarks — LiveBench (Apache-2.0)Elo — Arena Full attribution →