Models
Catalogue, pricing, usage, margin, cost advantage.
18 live models
Live models+2%
18
Revenue 30d+18.4%
$20.9K
Requests 30d+22.1%
19.9M
Tokens 30d+20.4%
53B
Avg margin+1.2%
46%
Open-source share
100%
- OpenPromote Qwen 2.5 72B vs Llama 3.1 70B for coding workloadsProduct · AI
Higher margin (61% vs 48%) with comparable benchmarks. Add a 'recommended for coding' tag in the catalogue.
- Re-price DeepSeek-R1 output to $1.80Pricing
Demand sensitive at this tier; competing API at $2.19. Margin still holds above 50%.
- OpenAudit low-margin image models in EU — below 25% thresholdPricing · Legal
Image generation models in EU show gross margin below the 25% floor. Review routing to higher-tier nodes or adjust regional pricing.
FAR AI vs centralized providers — $/1M tokens
~58% cheaperRequests by modality
Catalogue
All FAR AI models with pricing, usage, margin, latency, error.
| Model | Type | Status | Source | Input $/M | Output $/M | Requests 30d | Tokens | Revenue 30d | Margin | p95 | Err % |
|---|---|---|---|---|---|---|---|---|---|---|---|
Llama 3.1 405B Instruct Meta · llama-3.1-405b | text | live | Open | $1.2 | $1.8 | 1.9M | 5.4B | $7.2K | 40% | 2.1s | 2.3% |
DeepSeek-R1 DeepSeek · deepseek-r1 | reasoning | live | Open | $0.55 | $2.19 | 1.2M | 3.4B | $3.3K | 48% | 1.3s | 2.4% |
Llama 3.1 70B Instruct Meta · llama-3.1-70b | text | live | Open | $0.35 | $0.55 | 2.9M | 6.6B | $2.6K | 48% | 1.3s | 1.5% |
Mistral Large 2 Mistral · mistral-large-2 | text | live | Open | $0.8 | $1.2 | 795K | 2.1B | $1.9K | 48% | 1.3s | 3.3% |
Qwen 2 VL 72B Alibaba · qwen-2-vl-72b | vision | live | Open | $0.5 | $0.7 | 960.7K | 2.2B | $1.2K | 48% | 1.3s | 3.3% |
DeepSeek-V3 DeepSeek · deepseek-v3 | text | live | Open | $0.14 | $0.28 | 1.9M | 5.1B | $863.9 | 48% | 1.3s | 1.8% |
DeepSeek Coder V2 DeepSeek · deepseek-coder-v2 | coding | live | Open | $0.14 | $0.28 | 1.4M | 4.7B | $796.6 | 48% | 1.3s | 2.6% |
Qwen 2.5 72B Instruct Alibaba · qwen-2.5-72b | text | live | Open | $0.3 | $0.5 | 897K | 2.2B | $786.6 | 48% | 1.3s | 2.9% |
Mixtral 8x22B Instruct Mistral · mixtral-8x22b | text | live | Open | $0.45 | $0.75 | 520.6K | 1.1B | $587.1 | 48% | 1.3s | 3.4% |
QwQ 32B Preview Alibaba · qwq-32b | reasoning | live | Open | $0.18 | $0.32 | 659.5K | 1.7B | $360.5 | 54% | 720ms | 3.1% |
Llama 3.1 8B Instruct Meta · llama-3.1-8b | text | live | Open | $0.05 | $0.08 | 2.3M | 5.9B | $335.2 | 54% | 720ms | 1.3% |
Qwen 2.5 Coder 32B Alibaba · qwen-2.5-coder-32b | coding | live | Open | $0.12 | $0.2 | 865.6K | 2.2B | $305.6 | 54% | 720ms | 2.9% |
LLaVA 1.6 34B LLaVA · llava-1.6-34b | vision | live | Open | $0.4 | $0.6 | 205.2K | 674.3M | $297 | 48% | 1.3s | 3.4% |
Whisper Large v3 OpenAI · whisper-large-v3 | audio | live | Open | $0.04 | $0 | 551.9K | 1.8B | $60.3 | 54% | 720ms | 3.4% |
FLUX.1 [dev] Black Forest · flux-1-dev | image | live | Open | $0.03 | $0.04 | 742K | 2B | $57.3 | 48% | 1.3s | 3.1% |
FLUX.1 [schnell] Black Forest · flux-1-schnell | image | live | Open | $0.01 | $0.02 | 1.2M | 3.2B | $43.6 | 54% | 720ms | 2.9% |
BGE Large EN v1.5 BAAI · bge-large-en | embedding | live | Open | $0.02 | $0 | 609.5K | 1.8B | $26 | 54% | 720ms | 3.3% |
Nomic Embed Text v1.5 Nomic · nomic-embed-text | embedding | live | Open | $0.02 | $0 | 325.8K | 844.3M | $12.9 | 54% | 720ms | 3.5% |
Pixtral Large Mistral · pixtral-large | vision | beta | Open | $0.55 | $0.85 | 0 | 0 | $0 | 0% | 1.3s | 0% |
Stable Diffusion 3.5 Large Stability AI · sd-3.5-large | image | beta | Open | $0.02 | $0.03 | 0 | 0 | $0 | 0% | 1.3s | 0% |
1–20 of 22