Quality
How Meta: Llama 3.3 70B Instruct compares against every tracked model
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.
Instructthis #57/65
Quality Breakdown
Meta: Llama 3.3 70B Instruct on each evaluation feeding the index, against the tracked average
Bar = Meta: Llama 3.3 70B Instruct · grey marker = average across tracked models · scores ingested from Epoch AI, LiveBench
Arena Elo
Human preference rating from blind pairwise votes
Elo from blind pairwise votes, ingested from Arena
Ingested 4 Sep 2026.
Price & Cost
Published API pricing by token type
USD per 1M tokens at a 3:1 input:output blend. Live from OpenRouter, cross-checked against models.dev. Lower is better. Ranked against the 442 models priced above zero, which is the population the chart draws; excludes 29 published at no cost (free and promotional tiers).
Instructthis #71/442 · off scale
Context Window
Maximum input tokens accepted
Maximum input context from provider metadata via models.dev and OpenRouter. Higher is better.
Instructthis #375/476
Speed
Mean output throughput across hosting providers, over successful measurements from the last 48 hours
Mean output tokens per second across all tracked providers, over successful measurements from the last 48 hours. Derived from OpenRouter endpoint telemetry. Higher is better.
Instructthis #194/338 · below scale
Latency
Time to first token, mean across providers over successful measurements from the last 48 hours
Seconds to first streamed chunk, mean across tracked providers over successful measurements from the last 48 hours. Derived from OpenRouter endpoint telemetry. Lower is better.
Instructthis #110/338 · off scale
Providers
Hosts serving Meta: Llama 3.3 70B Instruct, with their own price and speed
| Provider | Input $/M | Output $/M | Blended | Speed | TTFT | Context |
|---|---|---|---|---|---|---|
| Groq | $0.59 | $0.79 | $0.64 | 228 | 0.2s | 131k |
| CoreWeave | $0.71 | $0.71 | $0.71 | 52 | 0.2s | 128k |
| Crusoe | $0.25 | $0.75 | $0.38 | 45 | 0.3s | 131k |
| Google · first party | $0.72 | $0.72 | $0.72 | 32 | 0.3s | 128k |
| Parasail | $0.22 | $0.50 | $0.29 | 28 | 0.6s | 131k |
| Together | $1.04 | $1.04 | $1.04 | 28 | 0.7s | 131k |
| SambaNova | $0.45 | $0.90 | $0.56 | 25 | 0.4s | 131k |
| Novita | $0.14 | $0.40 | $0.20 | 20 | 0.6s | 12k |
| AkashML | $0.20 | $0.52 | $0.28 | 19 | 0.8s | 131k |
| Cloudflare | $0.29 | $2.25 | $0.78 | 14 | 0.3s | 24k |
| DeepInfra | $0.10 | $0.32 | $0.15 | 14 | 0.6s | 131k |
| Nebius | $0.13 | $0.40 | $0.20 | 12 | 0.5s | 131k |
Per-host pricing and telemetry from OpenRouter. Speed is output tokens per second; TTFT is time to first token.
History
How price and quality have moved since we began tracking this model
Since 25 Jul 2026
| Known Good Index | 34 · based on 4 of 7 evaluations |
| Blended price | $0.15 per 1M · observed 4 Sep 2026 |
| Output speed | 44 tok/s · observed 4 Sep 2026 |
| Time to first token | 0.8s |
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.