AI Model Pricing Monitor

Direct-API and provider pricing for current frontier & value LLMs — USD per 1M tokens.

Data as of 2026-08-17 · generated 2026-09-08 · DeepSeek direct parsed from official docs · Nube live from console.nube.sh API · Fireworks/InferX curated
DeepSeek shown at peak rates; off-peak = 50% lower. DeepSeek peak hours: 01:00–04:00 & 06:00–10:00 UTC.

Cheapest input

Qwen3.7-Flash

Alibaba Qwen · $0.03 per 1M in

Cheapest output

DeepSeek V4 Flash

DeepSeek · $0.11 per 1M out

Largest context

GPT-5.6 Luna

OpenAI · 1.05M

Core requested

4 models

DeepSeek ×2 + Qwen ×2 · see table peak rates

Comparison table

Model	Tier	Context	Max out	Best price ($/M in / out)	Notes
DeepSeek V4 Flashcore DeepSeek	cheap	1M	384K	$0.07 / $0.11 via InferX · 4 providers	Efficiency-optimized MoE (284B/13B active). 1M ctx, thinking mode.
Nube-Choice Nube	cheap	1M	384K	$0.154 / $0.462 via Nube · 1 provider	Nube's auto-routing model. Currently routes to DeepSeek-V4-Flash (price tracks the routed model; subject to change as routing shifts).
DeepSeek V4 Procore DeepSeek	frontier	1M	384K	$1.32 / $3.96 via Direct API · 1 provider	DeepSeek's flagship. Same 1M ctx / 384K out, higher ceiling.
Qwen3.8-Maxcore Alibaba Qwen	frontier	1M	131K	$2.00 / $6.00 via Direct API · 1 provider	2.4T MoE (95B active), multimodal. Flagship of Qwen3.8 series.
Qwen3.7-Pluscore Alibaba Qwen	mid	1M	131K	$0.32 / $1.28 via Direct API · 1 provider	Cost-effective workhorse in the Qwen3.7 series.
Qwen3.7-Flash Alibaba Qwen	cheap	1M	131K	$0.03 / $0.13 via Direct API · 1 provider	Qwen's budget text tier, 1M ctx. Cheapest model on the board.
Qwen3.8-27B Alibaba Cloud (Qwen)	mid	0.262M	262K	$0.175 / $1.05 via Nube · 2 providers	Open-weight (Apache-2.0) dense VLM, 262K ctx. Served via Cloudflare/OpenRouter/Vercel.
GPT-5.6 Luna OpenAI	cheap	1.05M	128K	$0.2 / $1.20 via Direct API · 1 provider	OpenAI's cheapest GPT-5.6 tier. Fast, low-cost general model.
GPT-5.6 Terra OpenAI	mid	1.05M	128K	$2.00 / $12.00 via Direct API · 1 provider	Mid-tier GPT-5.6 (Terra). Balances cost and capability.
Claude Opus 5 Anthropic	frontier	1M	64K	$5.00 / $25.00 via Direct API · 1 provider	Anthropic's frontier reasoning model.
Claude Sonnet 5 Anthropic	mid	1M	64K	$2.00 / $10.00 via Direct API · 1 provider	Balanced mid-tier Claude for agents and tool use.
GLM-5.3 Zhipu AI	frontier	1M	—	$0.49 / $1.54 via Nube · 2 providers	Zhipu flagship coding/cyber model, 1M ctx.
GLM-5.3-Flash Zhipu AI	cheap	1M	—	$0.06 / $0.2 via Nube · 2 providers	Natively-multimodal (first in GLM-5 series), 320B/18B active, 1M ctx.
MiniMax M3 MiniMax	cheap	1M	—	$0.3 / $1.20 via Direct API · 1 provider	Multimodal foundation model; very cheap per token.
MiMo-V2.5 Xiaomi	cheap	1M	131K	$0.14 / $0.28 via Direct API · 1 provider	Xiaomi's omnimodal LLM (image/video/audio/text), 1M ctx.
Qwen3.7-Max Alibaba Qwen	frontier	1M	131K	$1.25 / $3.75 via Direct API · 1 provider	Previous Qwen flagship; promotional pricing active.
GPT-5.6 Sol OpenAI	frontier	1.05M	128K	$5.00 / $30.00 via Direct API · 1 provider	OpenAI's top GPT-5.6 tier.
Kimi K3 Moonshot AI	frontier	1M	—	$1.65 / $8.25 via Nube · 2 providers	Newest Kimi flagship, 1M ctx.

Cost calculator

Monthly requests

Avg input tokens / request

Avg output tokens / request

Cache-hit rate 0%

Price channel

DeepSeek timing

Model	Vendor	Est. monthly cost	Cost / 1M out	Tier

Recommendations

Cheapest across all providers (DeepSeek V4 Flash): InferX $0.07/$0.11 is the cheapest; Nube (recently raised to $0.286/$0.858) and direct are pricier.
Cheapest simple pick: Qwen3.7-Flash at $0.03/$0.13 (direct) — cheapest input on the board.
Frontier value: Qwen3.8-Max ($2/$6) is the value flagship; GLM-5.3 ($1.40/$4.40) a strong coding/cyber pick; Claude Opus 5 / GPT-5.6 Sol ($5/$25–$30) are the premium ceiling.
Cheapest multimodal: GLM-5.3-Flash — $0.06/$0.20 on Nube (Z.ai promo $0.075/$0.25 until Sep 9; list $0.15/$0.50) and MiMo-V2.5 ($0.14/$0.28).
Best code/agents value: GLM-5.3-Flash (Nube $0.06/$0.20) and GLM-5.3 (Nube $0.91/$2.86).
Largest context: DeepSeek V4, Qwen3.8, GPT-5.6, GLM-5.3 all offer ~1M-token context.

Similar models worth adding

Beyond the four requested models, these are the closest current comparables (already in the table above):

GPT-5.6 Luna (OpenAI) — cheap tier, $0.20/$1.20 direct.
MiniMax M3 — multimodal, ~$0.30/$1.20; among the cheapest flagships.
Kimi K3 (Moonshot) — flagship, $3/$15; older Kimi tiers ($0.06–$0.95 in) also strong value.
GLM-5.3 (Zhipu) — flagship coding/cyber, $1.40/$4.40.
GLM-5.3-Flash (Zhipu) — natively multimodal, $0.075/$0.25 (promo) / $0.15/$0.50 list.
Claude Sonnet 5 — balanced mid-tier, $2/$10.
Qwen3.7-Max — prior Qwen flagship on promo ($1.25/$3.75).
Grok / Mistral / Gemini — not yet priced here; add once live rates are confirmed.

Methodology & sources

Direct = the vendor's own API list price. Provider columns (Nube / Fireworks / InferX) = each provider's posted rate for that model. Prices are USD per 1M tokens; DeepSeek shown at peak with cache-hit input pricing — off-peak is 50% lower. Cache-hit discounts and batch pricing vary by vendor. Promotional prices are marked. Figures were hand-verified against official docs and provider pricing pages on the generation date.

Generated by generate_site.py · data model in models_data.py · live refresh + change alert via monitor.py (silent unless prices move).