GLM-4.6 API pricing

8 providers serve GLM-4.6. GLM-4.6 is an open-weight model (MIT) with 357B total parameters, 32B active per token (mixture-of-experts), up to 205K context. Prices below are per 1M tokens, cheapest input first (self-hosting math further down).

Refreshed 2026-08-23 · list prices from provider APIs

Input floor $0.400/1M at OpenRouter Output floor $1.75/1M at OpenRouter Find your workload winner →

Provider	Input $/1M	Output $/1M	Cache read $/1M	Context
OpenRouter	$0.400	$1.75output floor	-	203K	Use
Vercel AI Gateway	$0.450	$1.80	$0.110	200K	Use
Fireworks AI	$0.550	$2.19	-	203K	Use
Novita AI	$0.550	$2.20	$0.110	205K	Use
Baseten	$0.600	$2.20	-	-	Use
Together AI	$0.600	$2.20	-	200K	Use
Z.ai	$0.600	$2.20	$0.110	200K	Use
Cerebras	$2.25	$2.75	-	128K	Use

The cheapest way to run GLM-4.6

On an illustrative 70M-input / 30M-output monthly workload, today's lowest listed cost is roughly $80.50/month, at OpenRouter. Input-heavy, output-heavy and cache-heavy workloads can produce different winners. Use the live mix calculator below rather than combining floors from two different providers. On output-token cost alone, a busy self-hosted deployment can work out cheaper (~$0.659/1M output on 3× MI300X). See the math below. Run your own numbers in the breakeven calculator, or read the breakeven math.

Your actual monthly cost

Uses today's list prices

Workload preset Input tokens / month (millions) Output tokens / month (millions) Input served from cache %

Cache read uses the listed cache rate where available; otherwise it falls back to normal input price. Batch, write-cache, volume and negotiated discounts are excluded.

Or run it yourself: GLM-4.6 self-hosting economics

GLM-4.6 is open-weight (MIT), so the API price above competes with the GPU-hour market. 357B total parameters (MoE, ~32B active per token) needs roughly 491 GB VRAM at FP8 or 246 GB at INT4, KV-cache headroom included.

No single GPU fits GLM-4.6 at FP8 (491 GB needed), so it needs a multi-GPU node. Cheapest tracked option: 3× MI300X (192 GB each) at roughly $1.50/hr total (RunPod). Rent 3× MI300X at RunPod → INT4 quantization drops the requirement to 246 GB. Tune it in the calculator.

Breakeven estimate: a well-batched vLLM deployment on 3× MI300X ($1.50/hr) at ~1265 aggregate tok/s and 50% utilization works out to roughly $0.659 per 1M output tokens, versus $1.75 via the cheapest output-token API (OpenRouter). Self-hosting wins on cost if you can keep the GPU busy. Planning estimate: throughput varies with hardware, quantization, batch size and context. Tune it in the calculator.

Related models

Model	Cheapest in $/1M	Cheapest out $/1M	Providers
GLM-4.7	$0.400	$1.50	10
GLM-4.5 Air	$0.125	$0.450	7
GLM-5.2	$0.610	$1.98	6
GLM-4.5	$0.400	$1.60	6
GLM-5	$0.800	$2.56	5
GLM-5.1	$1.05	$3.50	4

FAQ

What is the cheapest API for GLM-4.6?

As of 2026-08-23, the lowest input price for GLM-4.6 is OpenRouter at $0.400 per 1M input tokens. The lowest output price is OpenRouter at $1.75 per 1M output tokens. The cheapest provider for a real workload depends on its input/output mix. The next-lowest input price is Vercel AI Gateway at $0.450, a 12% difference.

How much VRAM do you need to self-host GLM-4.6?

GLM-4.6 has 357B parameters, so plan for roughly 491 GB of VRAM at FP8 or 246 GB at INT4/AWQ, KV-cache headroom included. That exceeds a single GPU. A typical node is 3× MI300X.

← All models · GPU rental prices · Breakeven calculator · Get this data as JSON