Real-time Free vs Paid Index · Verified 2026
AI API Value & Status Leaderboard
The definitive live benchmark ranking foundational models by Intelligence per Dollar, token generation speed, and API operational status.
#1 Highest Value
Gemini 2.0 Flash
99/100 Value Score
#1 Fastest Streaming
Groq LPU
320 tokens / sec
#1 Coding Benchmark
Claude 3.5 Sonnet
95/100 Benchmark IQ
Cheapest Paid API
DeepSeek V3
$0.14 / 1M Input
Interactive Value Frontier (Logarithmic Scale)
Intelligence vs. Cost Quadrant Analysis
Logarithmic scale spreads ultra-cheap models across the canvas. Top-left green zone represents the high-IQ Pareto frontier.
Filter:
Cost-Efficient
Flagship
Open-Weights
Reasoning
Tip: Hover or tap any point to inspect specs
← Swipe horizontally to explore full quadrant →
DeepSeek V3DeepSeekCost-EfficientIQ: 89/100
State-of-the-art open economics, high-performance general intelligence
Blended (1M)$0.196
Input$0.14
Output$0.28
Throughput85 t/s
Latency (TTFT)380ms
Context128k
Real-Time Speed Benchmark (Artificial Analysis)
Live Token Streaming Speed (TPS) Arena
Simulate and visually compare real-time token streaming throughput and TTFT latency across models.
Workload:
Elapsed: 0.00s·Fastest: Llama 3.3 70B (Groq/Together) (3.4x)
Model 1310 tokens/sec
Latency: 140ms TTFTCluster: Groq LPU / Cerebras CS-3
Click 'Test Live Streaming Speed' to watch real-time generation...
Model 290 tokens/sec
Latency: 420ms TTFTCluster: Microsoft Azure NVIDIA H100
Click 'Test Live Streaming Speed' to watch real-time generation...
Rank by:
| Rank | Model & Provider | Tier / Allowance | Pricing (1M Tokens) | Speed | Value Score | API Status | Action |
|---|---|---|---|---|---|---|---|
| 1 | Gemini 2.0 Flash Ultra-fast multimodal streaming, vision & real-time chat | Free Tier 1,500 RPD on Google AI Studio | $0.100 in $0.40 out | 260 t/s streaming | 99/100 | Operational 99.98% Uptime | Inspect |
| 2 | DeepSeek R1 (Reasoning) PhD-level math, algorithmic reasoning, open deep thinking | Paid (Ultra Low) 5M tokens on signup | $0.550 in $2.19 out | 65 t/s streaming | 98/100 | Operational 99.98% Uptime | Inspect |
| 3 | DeepSeek V3 High-volume code refactoring & bulk scraping | Paid (Ultra Low) 5M promo tokens on signup | $0.140 in $0.28 out | 95 t/s streaming | 98/100 | Operational 99.98% Uptime | Inspect |
| 4 | Groq Cloud (Llama 3.3 70B) Sub-second real-time conversational voice agents | Free Tier 30 RPM / 6k TPM free on LPU | $0.590 in $0.79 out | 320 t/s streaming | 96/100 | Operational 99.98% Uptime | Inspect |
| 5 | GitHub Models (GPT-4o / Claude) Zero-cost developer prototyping & experimentation | Free Tier 15 RPM / 150 RPD in VS Code | $0.00 (Free) | 85 t/s streaming | 95/100 | Operational 99.98% Uptime | Inspect |
| 6 | GPT-4o Mini High-frequency customer chat & JSON extraction | Paid (Ultra Low) None (Requires API credits) | $0.150 in $0.60 out | 140 t/s streaming | 94/100 | Operational 99.98% Uptime | Inspect |
| 7 | Claude 3.5 Haiku Fast agentic sub-tasks & code syntax checking | Paid (Ultra Low) None | $0.800 in $4.00 out | 160 t/s streaming | 91/100 | Operational 99.98% Uptime | Inspect |
| 8 | Claude 3.5 Sonnet Gold-standard software engineering & autonomous agents | Paid (Flagship) None | $3.000 in $15.00 out | 75 t/s streaming | 89/100 | Operational 99.98% Uptime | Inspect |
| 9 | OpenAI o1-mini STEM reasoning, competitive coding, math algorithms | Paid (Mid Tier) None | $1.100 in $4.40 out | 60 t/s streaming | 88/100 | Operational 99.98% Uptime | Inspect |
| 10 | OpenAI GPT-4o General enterprise multimodal pipelines | Paid (Flagship) None | $2.500 in $10.00 out | 90 t/s streaming | 87/100 | Operational 99.98% Uptime | Inspect |
| 11 | Mistral Large 2 European multilingual compliance & GDPR datasets | Paid (Flagship) Experimentation tier available | $2.000 in $6.00 out | 80 t/s streaming | 86/100 | Operational 99.98% Uptime | Inspect |