Models & APIs / CoreWeave Forge Inference API pricing Verified 16 Nov 2026, 12:00 UTC Source coreweave.com/forge-pricing/inference-agents

CoreWeave Forge Inference API pricing

by CoreWeave Forge Inferenceper 1M tokens
$0.13 / 1M outputgpt-oss-20B (CoreWeave) · the cheapest model

CoreWeave Forge Inference API costs $0.03–$1.31 per 1M input tokens and $0.13–$3.96 per 1M output tokens. Output tokens cost several times more than input, so long answers drive the bill.

Cheapest output$0.13gpt-oss-20B (CoreWeave) · per 1M
Top model$3.96DeepSeek V4-Pro-0813 (CoreWeave) · per 1M output

Every model, per 1M tokens

USD list price, before batch or volume discounts
ModelInputOutputCached inputContext
DeepSeek V4.1-Flash (CoreWeave)$0.13$0.65$0.002—
IBM Granite 4.2 8B (CoreWeave)$0.06$0.15$0.015—
Qwen3.8 27B (CoreWeave)$0.40$3.00$0.15—
GLM 5.3 Flash (CoreWeave)$0.12$0.38$0.02—
DeepSeek V4-Pro-0813 (CoreWeave)$1.31$3.96$0.044—
NVIDIA Nemotron 3.5 Lightning (CoreWeave)$0.07$0.20$0.04—
DeepSeek V4-Flash-0731 (CoreWeave)$0.13$0.27$0.05—
GLM 5.2 (CoreWeave)$0.62$2.42$0.14—
MiniMax M3 (CoreWeave)$0.23$0.96$0.05—
Kimi K2.7 Code (CoreWeave)$0.60$3.00$0.12—
NVIDIA Nemotron 3 Ultra (CoreWeave)$0.50$2.15$0.10—
Kimi K2.6 (CoreWeave)$0.65$3.41$0.15—
Gemma 4 31B (CoreWeave)$0.10$0.34——
Gemma 4 26B A4B Instruct (CoreWeave)$0.10$0.30$0.05—
Qwen3.6 35B A3B (CoreWeave)$0.10$0.89$0.035—
DeepSeek V3.1 (CoreWeave)$0.50$1.50$0.20—
gpt-oss-20B (CoreWeave)$0.03$0.13——
gpt-oss-120B (CoreWeave)$0.03$0.17——
Llama 3.3 70B (CoreWeave)$0.71$0.71——
Llama 3.1 8B (CoreWeave)$0.22$0.22——
Bill estimate

What will you pay?

$5.27
per month on DeepSeek V4-Pro-0813 (CoreWeave)

Other model APIs

Compare all models →
Prices from coreweave.com/forge-pricing/inference-agents How we collect prices Some links are affiliate links; they never change the prices or the order shown. Disclosure