GLM 5.3 goes free at 12:00 PM IST · 24h window --:--:--

[ Evolvex · OpenAI-compatible gateway at edge.evlx.dev ]

One key. Every model. GLM 5.3, free.

Evolvex hands out free API keys to a single, OpenAI-compatible endpoint. Change one string — your base_url — and every SDK, agent and wrapper you already ship keeps working. GLM 5.3 runs open-house from 12:00 PM IST.

BASE_URL https://edge.evlx.dev/v1
38ms ttft 99.98% uptime 12:00 PM IST start €0 forever
GLM 5.3 FREE · 12 PM IST · 24H ·
Request — OpenAI compatible
▍ waiting to run…
Response — server-sent events
▍ idle
idle — the request will fire when you scroll to it
LIVE NOW Evolvex launch event · 12:00 PM IST · 24 hours

GLM 5.3, free for 24 hours.

Our flagship runs open-house with everything un-gated — full context, streaming, function calling. The only thing between you and the firehose is a deliberately minimal rate limit, so everyone gets a fair slice. Open-house runs until 12:00 PM IST tomorrow — full 200k context, streaming, function calling, all un-gated.

TIME LEFT IN THE EVENT
24HRS
:
00MIN
:
00SEC

100% of the event window left · free access ends at the next 12:00 PM IST

10 req/min — that's the whole cap 200k context, full streaming on no card, no email wall

Fair use applies · keys minted during the event keep working afterwards on the normal free tier

[ 01 ]

Why Evolvex

Four reasons to switch your base_url
/ 01

OpenAI-compatible

A strict drop-in. Point any OpenAI SDK, LangChain, Vercel AI or curl script at edge.evlx.dev/v1 and nothing else changes — same schemas, same errors, same streaming protocol.

/ 02

Streaming native

Server-sent events from the very first token. Time-to-first-token is measured in milliseconds, so agents, autocomplete and chat UIs feel instant — not eventually consistent.

/ 03

Edge-routed

Requests terminate at the nearest region and ride the fastest healthy upstream. One provider goes down? Failover happens mid-stream, before your users notice a thing.

/ 04

Free tier, real keys

Request a key, get real quota on the house — refreshed daily. We built Evolvex to give builders a working endpoint in under a minute, not a sales call.

[ 02 ]

Model index

One key unlocks every row
13 / 13 MODELS
ModelTypeContextStatus
gpt-4o chat · vision128k live
gpt-4o-mini chat · vision128k live
o3-mini reasoning128k live
claude-3.7-sonnet chat · reasoning200k live
claude-3.5-haiku chat200k live
deepseek-chat chat64k live
deepseek-reasoner reasoning64k live
llama-3.3-70b-instruct chat · open weights128k live
qwen-2.5-72b-instruct chat · open weights32k live
gemini-2.0-flash chat · vision1m live
mistral-large-2411 chat128k live
grok-2 chat131k beta
↳ no model matches that filter — try something broader.

▸ click any row to copy the model id straight into your clipboard

[ 03 ]

The network

Measured, not marketed
38ms
Median TTFT
first token, p50, 30d
99.98%
Uptime
rolling 90 days
12
Edge regions
anycast, auto-failover
2.4B
Tokens served
and counting
[ 04 ]

Pricing

There is only one plan
€0free, forever
no card · no trial clock · no asterisks
During the event: GLM 5.3, un-gated
24h from 12:00 PM IST · one light 10 req/min fairness cap
One free key per builder
minted instantly, yours to keep
50 requests / day, every model
quota refreshes at 00:00 UTC
Full streaming + function calling
nothing paywalled, nothing crippled
[ 05 ]

Claim your key

Takes about a minute

Ship tonight,
not next sprint.

Drop an email or a handle below and mint a key right here — it's automatically enrolled in the GLM 5.3 event (opens 12:00 PM IST). Manage quotas, logs and model toggles from your dashboard at edge.evlx.dev.

Your key — keep it secret, keep it shipping
▸ demo key for preview — redeem & manage the real one at edge.evlx.dev