EyesInAI
Switchyard
Benchmarks
Models
Learn
News
Benchmark your use case
⌘K
AI Benchmark Intelligence
Loading benchmark data…
EyesInAI
·
Loading live benchmark data
Home
Benchmark
Search
Refresh
Text
A
Loading models…
EyesInAI
·
Loading models…
All models
groq
meta-llama/llama-4-scout-17b-16e-instruct
Last tested Jun 25, 2026
Overall pass
100%
Avg latency
663 ms
Context
131k
Tools
Yes
Input $/1M
$0.11
Output $/1M
$0.34
Tests run
14
Passed
14/14
AI summary
Hardware to run it
Hosted:
$0.11 in / $0.34 out
per 1M
Self-host:
API only
Proprietary / weights not released — hosted API only.
Test results
⚡
Ping
Latency & availability — single-word reply
222 ms
🧮
Reasoning
Basic math reasoning — show work, give answer
297 ms
{ }
JSON Output
Structured output compliance — valid JSON with required keys
304 ms
💻
Code Gen
Python function generation with docstring
610 ms
🚀
Throughput
Token generation speed — 500-token long-form response
1346 ms
🔍
Context Recall
Retrieval from in-context data — 20-item list Q&A
Compare with another model
See the leaderboard
327 ms
🔧
Tool Use
Function/tool calling — get_weather invocation
655 ms
📝
Summarization
Real-world: distill a news article into 3 bullet points
741 ms
🏷️
Classification
Real-world: sentiment + category from customer review
357 ms
📄
Data Extraction
Real-world: pull structured fields from an invoice
553 ms
📋
Instruction Follow
Real-world: multi-rule compliance (5 sentences, no "the", etc.)
412 ms
✅
Format Compliance
IFEval-style: 4 bullets, keyword inclusion/exclusion, no preamble
2672 ms
🪡
Long-Context Needle
Find a 6-digit code buried in ~3.5k tokens of filler text
439 ms
🧠
Multi-Step Logic
BBH-lite: boolean expression + web of lies + object counting
343 ms