Models / Reka AI/ Reka Flash 3

Reka Flash 3

Reka AI · released Mar 11, 2025 · RekaAI/reka-flash-3

Input: text. Output: text.InputOutput
Type
Open weightsApache License 2.0
Params
20.9B
Context
66K

about 49K words of context

Our take

Written Aug 3, 2026

Reka Flash 3 is a 20.9-billion-parameter text model released in March 2025 with a permissive Apache licence. It handles up to 65,536 tokens in a single request and is available from two hosts at identical rates, so there is no price shopping between them.

Who should pick it

Pick this for permissive open-licence text work where commercial use, fine-tuning or redistribution matters. Use it when identical pricing across both hosts removes arbitrage effort. Skip it if you need benchmark scores, image or audio input, or guaranteed throughput.

The case for it

  • Apache 2.0 licence allows unrestricted commercial use, fine-tuning and redistribution.
  • Identical pricing across both tracked providers, so there is no arbitrage between hosts.

The case against it

  • No benchmark scores in our data, so quality is unmeasured.
  • Throughput is thin and uneven: only 14 tokens per second on one host, unverified on the other.
  • No disclosed active parameter count, so per-token efficiency is unknown.
00

How good is it?

We hold no score for this model.

We look for every model we track on every board we watch, and none of them has turned up Reka Flash 3 — so there is no intelligence, coding, agentic or writing score to show you, not a low one, none.

The boards we watch →

01

Can you run it yourself?

Fits in memory
weights load entirely on the card
Spills to system RAM
some weights offload; much slower
Too large
will not load even with offload
est
size is calculated; the verdict could change by 10%

Comfortable fit

A card many people ownFits in memory

GeForce RTX 4090 · 24 GB

Weights at Q4_K_M13.2 / 24 GBest
Spare memory7.7 GB spare
Usable context33K of 66K
Decode speed64 tok/sest

Room to spare. 7.7 GB spare means a 10% error in the size would not change the answer.

One step upFits in memory

GeForce RTX 5090 · 32 GB

Weights at Q4_K_M13.2 / 32 GBest
Spare memory15.7 GB spare
Usable context66K of 66K
Decode speed113 tok/sest

Room to spare. 15.7 GB spare means a 10% error in the size would not change the answer.

On a MacFits in memory

Apple M2 (10-core GPU) · 24 GB

Weights at Q4_K_M13.2 / 24 GBest
Spare memory2.9 GB spare
Usable context16K of 66K
Decode speed6 tok/sest

Room to spare. 2.9 GB spare means a 10% error in the size would not change the answer.

Your hardware
Checking your profile…

Memory use by level

Against a 24 GB card.

Q4_K_M
recommended
13.2 GBest
Fits in memory
Q5_K_M
15.5 GBest
Fits in memory
Q8_0
23.1 GBest
Spills to system RAM

All 3 sizes here are calculated, not measured. We hold no measured file for this model, so each size comes from the parameter count and every verdict above inherits that uncertainty.

Check against your own machine → · All 71 devices, with every size →

02

Or rent it from someone else

Cheapest published offer

Cheapest of 2 live listings. Picked at the widest standard context we hold, within one quantisation slice, so the numbers beside it are a price one host actually charges.

per 1M tokens
$0.10 in / $0.20 out
Context served
66K
Throughput
Not measured
Current provider offers with price, context and prompt-privacy answers
ProviderIn / out per 1M tokensContextThroughputTrains on promptsLogs promptsZero retention
OpenRouter$0.10 / $0.2066Knot measuredUnknownUnknownUnknown
Rekafp8$0.10 / $0.2066K54 tok/sNoNoConfirmed

Across the 2 listings we hold: 1 say they do not train on prompts, 0 say they do and 1 do not say. 1 appear in the zero-retention registry we check; the rest are unknown to us rather than confirmed either way.

What each host's API supports

From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.

Supported
Not supported
Not published
host gave no parameter list
API features per host
ProviderTool callingJSON outputStrict schema
OpenRouter
Rekafp8

Tool calling: 0 of 2 listings say yes, 2 say no. JSON output: 0 of 2 listings say yes, 2 say no. Strict schema: 2 of 2 listings say yes.

03

When we formed this view

Dates behind this page

Jul 26, 2026ListedListed on LLMapfirst indexed by our pipeline
Mar 11, 2025AnnouncedReka Flash 3 announced by Reka AI

Prices last checked 35h ago

What we do not know about this model yet

  • We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
  • No board we watch has turned up a score, so we hold no quality figures at all.
  • Nothing we hold says whether an endpoint streams, so we do not show it either way.
  • 1 of 2 listings do not say whether they train on prompts.
  • We hold no cached-input rate for any of its listings.
04

Licence and identifiers

What the licence allowsApache License 2.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.

Licence

Apache License 2.0

permissiveCommercial use allowed

Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.

Identifiers

Architecture
Dense
Modality record
text->text
Catalogue slug
rekaai-reka-flash-3

Machine-readable model card (omc.json) →

Something wrong on this page? Tell us