LLMHOSTING.AI the compute price terminal

Kimi K2 API pricing

11 providers serve Kimi K2. Kimi K2 is an open-weight model (Modified MIT) with 1058B total parameters, 32B active per token (mixture-of-experts), up to 262K 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 $0.500/1M at DeepInfra Output floor $2.00/1M at DeepInfra and Hyperbolic Find your workload winner →
Provider Input $/1M Output $/1M Cache read $/1M Context
DeepInfra $0.500 $2.00output floor - 131K Use
Vercel AI Gateway $0.550 $2.20 - 131K Use
OpenRouter marketplace quote $0.570 $2.30 - 131K Use
Fireworks AI $0.600 $2.50 - 131K Use
Baseten $0.600 $2.50 - - Use
Moonshot $0.600 $2.50 $0.150 131K Use
W&B Inference $0.600 $2.50 - 128K Use
Novita AI $0.600 $2.50 - 262K Use
Groq $1.00 $3.00 $0.500 262K Use
Together AI $1.00 $3.00 - - Use
Hyperbolic $2.00 $2.00 - 131K Use

The cheapest way to run Kimi K2

On an illustrative 70M-input / 30M-output monthly workload, today's lowest listed cost is roughly $95.00/month, at DeepInfra. 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 (~$1.76/1M output on 8× 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

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: Kimi K2 self-hosting economics

Kimi K2 is open-weight (Modified MIT), so the API price above competes with the GPU-hour market. 1058B total parameters (MoE, ~32B active per token) needs roughly 1455 GB VRAM at FP8 or 728 GB at INT4, KV-cache headroom included.

No single GPU fits Kimi K2 at FP8 (1455 GB needed), so it needs a multi-GPU node. Cheapest tracked option: 8× MI300X (192 GB each) at roughly $4.00/hr total (RunPod). Rent 8× MI300X at RunPod → INT4 quantization drops the requirement to 728 GB. Tune it in the calculator.
Breakeven estimate: a well-batched vLLM deployment on 8× MI300X ($4.00/hr) at ~1265 aggregate tok/s and 50% utilization works out to roughly $1.76 per 1M output tokens, versus $2.00 via the cheapest output-token API (DeepInfra and Hyperbolic). 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

ModelCheapest in $/1MCheapest out $/1MProviders
Kimi K2 Thinking $0.600 $1.20 9
Kimi K2.5 $0.500 $2.80 8
Kimi K2.6 $0.950 $4.00 6
Kimi K2.7 Code $0.670 $3.40 4
Kimi K3 $3.00 $15.00 3
MoonshotAI Kimi Latest $2.00 $5.00 2

FAQ

What is the cheapest API for Kimi K2?

As of 2026-08-23, the lowest input price for Kimi K2 is DeepInfra at $0.500 per 1M input tokens. The lowest output price, $2.00 per 1M tokens, is likewise shared by DeepInfra and Hyperbolic. The cheapest provider for a real workload depends on its input/output mix. The next-lowest input price is Vercel AI Gateway at $0.550, a 10% difference.

How much VRAM do you need to self-host Kimi K2?

Kimi K2 has 1058B parameters, so plan for roughly 1455 GB of VRAM at FP8 or 728 GB at INT4/AWQ, KV-cache headroom included. That exceeds a single GPU. A typical node is 8× MI300X.

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