jishie

Kimi K3

hyper/kimi-k3 · family kimi-k3 · in: text, image → out: text · price provider-declared (via models.dev). Capabilities: tools reasoning text image.

$6.53
Blended $/1M (3:1)
$3.27
Input $/1M
$16.33
Output $/1M
$19.60
1M in + 1M out
$0.3266 −90%
Cache read $/1M
#6904 of 7,708
Cheapest rank
Provider
Charm Hyper (hyper)
Kind
text (outputs text)
Context
1,048,576 tokens
Max output
16,000 tokens
Capabilities
tools reasoning text image

Same model, other providers

“Kimi K3” is offered by 70 providers — cheapest-effective first. SenseNova (China) is the cheapest at $0.00 blended.

ProviderInputOutputBlendedContext
SenseNova (China)freefree$0.001M
Umans AI Coding Planfreefree$0.001M
Volcengine Ark Coding Planfreefree$0.001M
Kimi For Coding (kimi.com)freefree$0.001M
SCNet Token Planfreefree$0.001M
Nvidiafreefree$0.001M
Kenarifreefree$0.001M
Kimi For Coding (kimi.ai)freefree$0.001M
CrofAI$2$8$3.501M
NanoGPT$2$10$4.001M
ainetcafe$2.1$10.5$4.20262k
Vancine$2.4$12$4.801M
Charm Hyper (this)$3.27$16.33$6.531M

Call it

jishie indexes & prices models; it does not proxy inference.

Most providers are OpenAI-compatible — point base_url at Charm Hyper and pass this model id:

curl $BASE_URL/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" \
  -d '{"model":"kimi-k3","messages":[{"role":"user","content":"hello"}]}'

Route programmatically with MCP find_model / get_model, or fetch this record free (no wallet):

← all Charm Hyper models · model index · JSON