Bitcoin puzzle transaction ~32 BTC prize to who solves it

[kTimesG]

Nobody says they broke 3 different mathematical structures. There is definitely no one who claims to have done such a thing.

We already explained a million times why, formally, the prefix theory would break 3 different mathematical constructs if true.

What I proposed doesn’t do it, you assume too much, you say, 'the theory doesn’t work'. Which one? Which of the 20? Just me alone, I have 2 approaches, WP has another, Bibilgin has another, NoMachine... and I could go on. That’s why your words don’t hold value.

Guys, please hold on 5 minutes to get my popcorn ready!

I think FUD is when someone says and proves this: "look, I can find a key faster than what the crypto framework guarantees". You will definitely be seen on TV if that day ever comes. No matter what theory you invented, as long as it works. Until then: details about this or that don't even matter.

Reality-check wake-ups are FUDs maybe for you only, since you should be the only ones worried that what you believe turns out to actually not work.

[Akito S. M. Hosana]

Why doesn’t anyone have a script to search for WIF patterns? Is it too hard?    

Not only is it difficult—it's impossible. Let's say we know that for puzzle 69, the WIF will start with KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q — 19 characters are missing. Somehow, you might manage to guess 11 or 12 characters... but 19 is impossible. Here's a useless script you can test :

Code:

import sys
import os
import time
import multiprocessing
from multiprocessing import cpu_count, Event, Value, Process
import numpy as np
from numba import njit, prange
import secp256k1 as ice

# Configuration - puzzle 68
START_WIF = "KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qd7sDG4F"
MISSING_CHARS = 52 - len(START_WIF)
TARGET_HEX = "e0b8a2baee1b77fc703455f39d51477451fc8cfc"
TARGET_BINARY = bytes.fromhex(TARGET_HEX)
BATCH_SIZE = 100000

# Global variables
STOP_EVENT = Event()
KEY_COUNTER = Value('i', 0)
START_TIME = Value('d', 0.0)
CHARS = np.frombuffer(
    b"123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz",
    dtype=np.uint8
)
START_BYTES = START_WIF.encode('ascii')  # Precompute this

@njit(cache=True, parallel=True)
def numba_generate_batch(start_bytes, miss, batch_size, chars):
    results = np.empty((batch_size, len(start_bytes) + miss), dtype=np.uint8)
    char_len = len(chars)
    for i in prange(batch_size):
        # Copy the fixed prefix
        results[i, :len(start_bytes)] = start_bytes
        # Generate random suffix with indices within bounds
        for j in range(miss):
            results[i, len(start_bytes)+j] = np.random.randint(0, char_len)
    return results

def generate_batch(batch_size):
    indices = numba_generate_batch(
        np.frombuffer(START_BYTES, dtype=np.uint8),
        MISSING_CHARS,
        batch_size,
        CHARS
    )
    return [START_BYTES + CHARS[indices[i, -MISSING_CHARS:]].tobytes() 
            for i in range(batch_size)]

def check_private_key_batch(target_binary):
    local_counter = 0
    
    while not STOP_EVENT.is_set():
        # Generate a batch of keys
        wif_batch = generate_batch(BATCH_SIZE)
        local_counter += BATCH_SIZE
        
        # Update global counter
        with KEY_COUNTER.get_lock():
            KEY_COUNTER.value += BATCH_SIZE
        
        # Process the batch
        for wif_bytes in wif_batch:
            if STOP_EVENT.is_set():
                break
                
            try:
                private_key_hex = ice.btc_wif_to_pvk_hex(wif_bytes.decode('ascii'))
                dec = int(private_key_hex, 16)
                ripemd160_hash = ice.privatekey_to_h160(0, True, dec)
                
                if ripemd160_hash == target_binary:
                    handle_success(dec)
                    return
                    
            except:
                continue
    
    # Add any remaining keys if we were interrupted
    with KEY_COUNTER.get_lock():
        KEY_COUNTER.value += local_counter % BATCH_SIZE

def handle_success(dec):
    t = time.ctime()
    wif_compressed = ice.btc_pvk_to_wif(dec)
    elapsed = time.time() - START_TIME.value
    
    with open('winner.txt', 'a') as f:
        f.write(f"\n\nMatch Found: {t}")
        f.write(f"\nPrivatekey (dec): {dec}")
        f.write(f"\nPrivatekey (hex): {hex(dec)[2:]}")
        f.write(f"\nPrivatekey (wif): {wif_compressed}")
        f.write(f"\nTotal keys checked: {KEY_COUNTER.value:,}")
        f.write(f"\nAverage speed: {KEY_COUNTER.value/elapsed:,.0f} keys/sec")
    
    STOP_EVENT.set()
    print(f"\n\033[01;33m[+] BINGO!!! {t}\n")

if __name__ == '__main__':
    os.system("clear")
    print(f"\033[01;33m[+] {time.ctime()}")
    print(f"[+] Missing chars: {MISSING_CHARS}")
    print(f"[+] Target: {TARGET_HEX}")
    print(f"[+] Starting WIF: {START_WIF}")
    print(f"[+] Cores: {cpu_count()}")
    
    # Initialize START_TIME
    START_TIME.value = time.time()
    
    try:
        os.nice(-20)
        import psutil
        p = psutil.Process()
        p.cpu_affinity(list(range(cpu_count())))
    except:
        pass

    workers = []
    for _ in range(cpu_count()):
        p = Process(target=check_private_key_batch, args=(TARGET_BINARY,))
        p.start()
        workers.append(p)
    
    try:
        while not STOP_EVENT.is_set():
            time.sleep(1)
            current_count = KEY_COUNTER.value
            elapsed = max(time.time() - START_TIME.value, 0.001)
            speed = current_count / elapsed
            sys.stdout.write(f"\r[+] Speed: {speed:,.0f} keys/sec | Total: {current_count:,} keys")
            sys.stdout.flush()
    except KeyboardInterrupt:
        STOP_EVENT.set()
        print("\n[!] Stopping workers...")
    
    for p in workers:
        p.join()
    
    print(f"\nSearch completed. Final count: {KEY_COUNTER.value:,} keys")

• Sat Apr 12 12:23:07 2025
• Missing chars: 12
• Target: e0b8a2baee1b77fc703455f39d51477451fc8cfc
• Starting WIF: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qd7sDG4F
  
• Cores: 12
• Speed: 301,173 keys/sec | Total: 78,100,000 keys
• BINGO!!! Sat Apr 12 12:27:27 2025
• Speed: 300,400 keys/sec | Total: 78,200,000 keys

Search completed. Final count: 78,200,000 keys

Match Found: Sat Apr 12 12:27:27 2025
Privatekey (dec): 219898266213316039825
Privatekey (hex): bebb3940cd0fc1491
Privatekey (wif): KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qd7sDG4F2sdMtzNe8y2U
Total keys checked: 78,100,000
Average speed: 300,812 keys/sec


You need a remote viewer to tell you exactly 7 characters (here is d7sDG4F) that are missing.  

I tested this—amazing! It solved 12 characters in 8 minutes. Is there a C++ version of this? I can only imagine what a script that's 200x faster could do.   

[bibilgin]

"Written by a 50-year-old retired man who is bored at home..." Bla bla bla... lol

I wrote you the facts above. Don't be so afraid of the facts. lol
Now go eat the corn and fish. You can come back when you are useful for something else.

As long as you don't respect, you can't be respected.
It doesn't matter what you respect or who you respect. What matters is to RESPECT.

You don't need to give away your biography. In your opinion, there are thousands of people in the city I live in.

Retired, over 50, spending their time watering the garden, playing backgammon, old men who are useless but disrespectful to everyone's opinion and don't care about their thoughts. lol

[nomachine]

I tested this—amazing! It solved 12 characters in 8 minutes. Is there a C++ version of this? I can only imagine what a script that's 200x faster could do.    

I have a script in C++, but it's not 200x faster—it doesn't use AVX2 hashing or the JLP SECP256K1.
It uses OpenSSL.

But what are you going to do with it?

To generate the number of possible combinations between "ecrA1gh" and "kW1gt2H" (7 missing characters for puzzle 69) using the Base58 character set, there are approximately 1.54 trillion combinations.

• WIFRotator
• Starting WIF: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q
• Middle range: ecrA1gh to kW1gt2H
• Missing chars: 12
• Target: 61eb8a50c86b0584bb727dd65bed8d2400d6d5aa
• Cores: 12
• Initial middle: f9FtYtY
• New middle section activated: ipNX8dr
• Speed: 16,2 MKeys/sec | Total: 6,404 MKeys
• New middle section activated: hGnuha8
• Speed: 16,2 MKeys/sec | Total: 6,304 MKeys
• New middle section activated: gxN4jE6
• Speed: 15,7 MKeys/sec | Total: 6,204 MKeys
• New middle section activated: fT9yPyt
• Speed: 15,9 MKeys/sec | Total: 6,404 MKeys
• New middle section activated: hM42ZDc
• Speed: 16,2 MKeys/sec | Total: 6,304 MKeys
• New middle section activated: h5KwaZ5
• Speed: 16,2 MKeys/sec | Total: 6,300 MKeys
• New middle section activated: jSaH3oc
• Speed: 16,2 MKeys/sec | Total: 6,301 MKeys
• New middle section activated: iHgzCdB

This means your script will rotate through ~1.54 trillion different middle sections (each taking about 1–8 minutes), while brute-forcing the last 12 characters for each one.  

[kTimesG]

"Written by a 50-year-old retired man who is bored at home..." Bla bla bla... lol

I wrote you the facts above. Don't be so afraid of the facts. lol
Now go eat the corn and fish. You can come back when you are useful for something else.

As long as you don't respect, you can't be respected.
It doesn't matter what you respect or who you respect. What matters is to RESPECT.

You don't need to give away your biography. In your opinion, there are thousands of people in the city I live in.

Retired, over 50, spending their time watering the garden, playing backgammon, old men who are useless but disrespectful to everyone's opinion and don't care about their thoughts. lol

So resorting to hallucinations about how you view me in your mind represents your defense on how you didn't broke the crypto, but still you broke the crypto? If you claim you didn't break the crypto, then why are you bothering chasing prefixes, since it should not help (otherwise - it means you broke the crypto)? Do you understand the logical contradiction?

Last time you called me too young, now you call me too old. I guess this is how you also compute your prefix search ranges. It fits well with your guesses so far - both wrong, without any recorded success.

[WanderingPhilospher]

We do not need to conduct empirical testing because of that. But it you want to, you can use my 67 proofs to do so, which will also disprove the prefix theory.

How did it disprove anything?

Were there multiple h160 matches that were back to back to back or something? Did I miss that?

I dunno about others, but I have been talking about averages...and I think your public data set, kind of proves what I was saying weeks back.

Code:

First Run:
- Average difference: 282602011632656
- Smallest difference: 194903573833
- Largest difference: 1946984192923367

Second Run (Excluding Smallest and Largest Differences):
- Average difference: 281241799946404

For reference, 2^48 = 281474976710656

So I do not know what it disproved, from my point of view, just the opposite really, that over x amount size range, an average starts to form for difference/distances apart.
That's all I've seen anyone say on the matter. No one claims to have made finding the key faster, or a better method/way. I think some are just trying to improve their odds, with little fire power, by using averages, versus a lot of fire power and random + sequential. That's all.
You can say you do not see how it helps and I would counter with, how does it hurt??

Also, in general, for others on the forum, so what if people "collect" or search for prefixes? How does it hurt? If someone is randomly searching the range and find a prefix, cool...but you know what else could happen while searching for a prefix...they could find the full address. No harm, no foul.

[bibilgin]

.lol  - Let's repeat. Bla bla bla...

Does anyone really understand this person?
I'm sure he reads what he wrote over and over again.

KTim - I am me, you are you.
- So what next?
KTim - Blue is beautiful.
Then?
KTim - Green is green.
Then you ask if there is anything?
KTim - The world is round. he says.

This drug head. lol

[kTimesG]

.lol  - Let's repeat. Bla bla bla...

Does anyone really understand this person?
I'm sure he reads what he wrote over and over again.

KTim - I am me, you are you.
- So what next?
KTim - Blue is beautiful.
Then?
KTim - Green is green.
Then you ask if there is anything?
KTim - The world is round. he says.

This drug head. lol

Someone please let me know when this user stops with the continuous insults, while he's asking for respect.

I remembered why he got ignored several times, I won't do that mistake again. I guess logical inference, a thing that exists for several thousands years since ancient philosophers, is not really something he can grasp in 2025.

[uyrova818]

i found a key that matches the first 7 characters of the target prefix (61eb8a5...) and the last 3 characters (...5aa). will this work?

up. I'm really curious about that

[bibilgin]

Someone please let me know when this user stops with the continuous insults, while he's asking for respect.

I remembered why he got ignored several times, I won't do that mistake again. I guess logical inference, a thing that exists for several thousands years since ancient philosophers, is not really something he can grasp in 2025.

The one who ignored was just you. lol

Let's go back to the beginning;
I told you in the first messages, you look down on people. You are being disrespectful. If you have knowledge, share it or have a logical conversation.

If you are tired of people's absurd, sometimes stupid questions without doing research, don't answer. If you have to write, you write a proper article.

There are so many people here. I have not written or insulted anyone in any way.
(Except you.) Because you see yourself as a smart person, a professor or even a philosopher who thinks that only he knows how to use artificial intelligence.

You do not RESPECT people's ideas, thoughts, words.

If you wish, read it again and again from the point we first started the discussion.

Yes, the year is 2025. Although years pass, there are important things even in the year 2505.
Some virtues are important. RESPECT - THOUGHT - IDEAS - LOVE - TOLERANCE... these are important.

[papiro08]

PubAddress: 19vkiEajcG4MNPMThVK2Kxog3vxuNhRaY4
Priv (WIF): p2pkh:KzHF5EkDnaLv7sGvs4orAFZr1QfuDZiCTL7Nc5A1uBFsxizNBcZC
Priv (HEX): 0x5B5458C8FB767377A02A6D1DD5687DD6B8F94A88500C8B9364CD9F911015C92D
[107708.64 Mkey/s][GPU 107488.26 Mkey/s][Total 2^47.59][Prob 100.0%][99% in 00:00:00][Found 225]
PubAddress: 19vkiEaj73nZvS4ePfXWuTwCNaZsEmhkEn
Priv (WIF): p2pkh:KwszBwcN3x3TqcDtj5WXfENUekhRzr181WRSwH29goMtoeXNxhP3
Priv (HEX): 0x13B0CAEEDEB14E00A2D5F62BA26B69F18AA1A7B83246526A74A981ADFE5D8EE3
[107708.75 Mkey/s][GPU 107488.21 Mkey/s][Total 2^47.59][Prob 100.0%][99% in 00:00:00][Found 226]
PubAddress: 19vkiEaj59ivaKeicDquPCHWYCDFBHM5nQ
Priv (WIF): p2pkh:L3imeAmxMhvqS5hRF19Np8jqkwLUBFXkpxeE58dQopyNoDr5yshJ
Priv (HEX): 0xC1F94716160E1E2743510310DFCBEDC21D493B65A8D5869753C668156B734559
[107607.04 Mkey/s][GPU 107386.73 Mkey/s][Total 2^47.60][Prob 100.0%][99% in 00:00:00][Found 227]
PubAddress: 19vkiEajn7rZH3kVdsUMZvb3nRDV7Mx8Lm
Priv (WIF): p2pkh:L2ZtSz9Kn5NwAp782xFq4kQJCA9c4cJXXkcWi38tAqqqJHrPG8k8
Priv (HEX): 0x9F911888F12591DD809C71A200A7C067424A6DBFDFD8AF8BF61151F3D20FA891
[107555.62 Mkey/s][GPU 107335.30 Mkey/s][Total 2^47.60][Prob 100.0%][99% in 00:00:00][Found 228]
PubAddress: 19vkiEajzfiRWYe7TC4YbgCvBhHay2Eux2
Priv (WIF): p2pkh:L5GUymChqHU3pGXPRWXiiZUXi2o7BGfHYUWcMzJUR7JsfnF4nN4N
Priv (HEX): 0xF02017BDE5C94B86329EB8C8B98C7F9DEE1100113E276B2104340CDD0C1CC221
[107555.35 Mkey/s][GPU 107334.68 Mkey/s][Total 2^47.61][Prob 100.0%][99% in 00:00:00][Found 229]
PubAddress: 19vkiEajw8F2yPzxN7QyCKiCXuzH2SKDja
Priv (WIF): p2pkh:KxJjscHJiv2CmLysNs5UftChePyBcwPTG3e8F8Y3bTS5CYZ8Q7tU
Priv (HEX): 0x206CB04BB7E859E0FB1A6DBC6DFC4B453FE9190EA5506BFEC2B44776F0457565
[107606.35 Mkey/s][GPU 107385.67 Mkey/s][Total 2^47.61][Prob 100.0%][99% in 00:00:00][Found 230]
PubAddress: 19vkiEajoHda5j3zhrxzhRNHNLM7z9dWEF
Priv (WIF): p2pkh:KycJha9gMATYrQ9jcrmbP7P7UrXQT3UFayCT7z6TvtbdJbfMuCxQ
Priv (HEX): 0x474C6EF455C88A7A2A6C87933EA456AF067F11533AB0F1805397F257A2C1B0E2
[107709.20 Mkey/s][GPU 107488.38 Mkey/s][Total 2^47.62][Prob 100.0%][99% in 00:00:00][Found 231]

[bibilgin]

PubAddress: 19vkiEajoHda5j3zhrxzhRNHNLM7z9dWEF
Priv (WIF): p2pkh:KycJha9gMATYrQ9jcrmbP7P7UrXQT3UFayCT7z6TvtbdJbfMuCxQ
Priv (HEX): 0x474C6EF455C88A7A2A6C87933EA456AF067F11533AB0F1805397F257A2C1B0E2
[107709.20 Mkey/s][GPU 107488.38 Mkey/s][Total 2^47.62][Prob 100.0%][99% in 00:00:00][Found 231]

Searches below 44 bits have no meaning.

Keep that in mind, my friend.

[papiro08]

I imagine that with 2000 GPUs you will have tons of much longer prefixes.

[WanderingPhilospher]

This thread is increasingly resembling a helpdesk for beginners. I can understand that there's a lot to learn, and we learn something new every day, but a little personal research would help.

I see people here and there looking for miracle prefixes, constantly changing programs, or creating 36 different ones to find a key quickly, but there are no miracles at this stage.

No algorithm, no method currently available, allows you to find a private key quickly. The best strategy if you're short on resources is to let a good program run and hope you're lucky enough to find the key, because the key is patience.

Use BSGS cuda, Kangaroo by RetiredCoder for public keys, and Cyclone or KeyQuest by Benjade if you only have CPUs that perform very well; these are currently the best. Forget vanity search, mutagen, and other pseudo-scripts; you'll be wasting your time, and time is money.

CZ said, "If you can't hold, you'll never be rich." I think this applies not only to the stock market, but to patience in general.

Oh no, what do us poor old souls use that have GPUs, for keys with no known public key??!! No fork of VanitySearch? I am short on resources but since you only stated 4 programs that were good, and none to work with GPU and brute force, I am lost. I need to find a good program for GPU and let it run and run and run until Sally and Dick are both tired. Please advise. Thank you!!!

[cctv5go]

I tested this—amazing! It solved 12 characters in 8 minutes. Is there a C++ version of this? I can only imagine what a script that's 200x faster could do.    

I have a script in C++, but it's not 200x faster—it doesn't use AVX2 hashing or the JLP SECP256K1.
It uses OpenSSL.

But what are you going to do with it?

To generate the number of possible combinations between "ecrA1gh" and "kW1gt2H" (7 missing characters for puzzle 69) using the Base58 character set, there are approximately 1.54 trillion combinations.

• WIFRotator
• Starting WIF: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q
• Middle range: ecrA1gh to kW1gt2H
• Missing chars: 12
• Target: 61eb8a50c86b0584bb727dd65bed8d2400d6d5aa
• Cores: 12
• Initial middle: f9FtYtY
• New middle section activated: ipNX8dr
• Speed: 16,2 MKeys/sec | Total: 6,404 MKeys
• New middle section activated: hGnuha8
• Speed: 16,2 MKeys/sec | Total: 6,304 MKeys
• New middle section activated: gxN4jE6
• Speed: 15,7 MKeys/sec | Total: 6,204 MKeys
• New middle section activated: fT9yPyt
• Speed: 15,9 MKeys/sec | Total: 6,404 MKeys
• New middle section activated: hM42ZDc
• Speed: 16,2 MKeys/sec | Total: 6,304 MKeys
• New middle section activated: h5KwaZ5
• Speed: 16,2 MKeys/sec | Total: 6,300 MKeys
• New middle section activated: jSaH3oc
• Speed: 16,2 MKeys/sec | Total: 6,301 MKeys
• New middle section activated: iHgzCdB

This means your script will rotate through ~1.54 trillion different middle sections (each taking about 1–8 minutes), while brute-forcing the last 12 characters for each one.  

Can you share this program? To try your luck, at least closer than the end of the universe.What is the random probability that these seven characters in the middle are different?

[WanderingPhilospher]

Somebody let me know if I find the full address before Bram

t.me/puzzleprefixes

Those are coming in real-time. But I may be asleep when the real deal Holyfield address is found...so, wake me, if need be, LOL!

Increased the minimum match to 44 bits as 40 bits was creating too many results being sent to channel.

Wake me up LOL!

Changed it again to a minimum of 48 bits...still too many at 44 bits.

Let me know if anyone wants to donate to the Beat Bram fund

[brainless]

@nomachine
Have you tried to make cyclone version with load of list hash160 instead of 1 h160
And speed effects ?
If still not worked...
Could you try to develop this version for load of hash160 list

[kTimesG]

Let's go back to the beginning;
I told you in the first messages, you look down on people. You are being disrespectful. If you have knowledge, share it or have a logical conversation.

Logical conversation:

You: "No one said they broke 3 math structures. Oh, btw, look, I can find John without knocking 10k doors..."
Me: "You are contradicting yourself."
You: "Drug head lol"

Unless someone is on weed, IDK how can they ever engage intn such a high-level argumentative logic.

I wonder why none of the prefix believers fails to answer this simple question:

Assuming the prefix theory works, it automatically implies a weakness in the cryptographic guarantees of Bitcoin, and more generally automatically implies that the elliptic curve cryptographic assumptions are broken, and also the cryptographic security of SHA256 and/or RIPEMD-160 are also broken.

Yielding BTC as a 0$ asset. Hence, why would you try to solve the puzzle?

If so, why the hell aren't you guys publish a paper and wait for the academic gratitude? It would worth infinitely more than some stupid satoshis... RIGHT ?!?

Maybe because you know deep inside that you're simply just chasing ghosts of the statistical fluctuations of something that is not distinguishable from randomness? Or are you afraid that you'll need to call all the academic peers who'd dismiss your paper in 5 seconds, as people who obviously "don't understand probabilities" and have zero clues about math?

Up.

[Dragon914421]

I am silently watching and learning from all you guys that have been in this space for quite some long time and doing the puzzle, but again and again people say you need to find more than one or that. What exactly does that mean? Is it if someone has multiple addresses with private key in 69bit range that starts with 19vkiEajf, does it help to achieve goal faster

[JimsR]

I tested this—amazing! It solved 12 characters in 8 minutes. Is there a C++ version of this? I can only imagine what a script that's 200x faster could do.    

I have a script in C++, but it's not 200x faster—it doesn't use AVX2 hashing or the JLP SECP256K1.
It uses OpenSSL.

But what are you going to do with it?

To generate the number of possible combinations between "ecrA1gh" and "kW1gt2H" (7 missing characters for puzzle 69) using the Base58 character set, there are approximately 1.54 trillion combinations.

• WIFRotator
• Starting WIF: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q
• Middle range: ecrA1gh to kW1gt2H
• Missing chars: 12
• Target: 61eb8a50c86b0584bb727dd65bed8d2400d6d5aa
• Cores: 12
• Initial middle: f9FtYtY
• New middle section activated: ipNX8dr
• Speed: 16,2 MKeys/sec | Total: 6,404 MKeys
• New middle section activated: hGnuha8
• Speed: 16,2 MKeys/sec | Total: 6,304 MKeys
• New middle section activated: gxN4jE6
• Speed: 15,7 MKeys/sec | Total: 6,204 MKeys
• New middle section activated: fT9yPyt
• Speed: 15,9 MKeys/sec | Total: 6,404 MKeys
• New middle section activated: hM42ZDc
• Speed: 16,2 MKeys/sec | Total: 6,304 MKeys
• New middle section activated: h5KwaZ5
• Speed: 16,2 MKeys/sec | Total: 6,300 MKeys
• New middle section activated: jSaH3oc
• Speed: 16,2 MKeys/sec | Total: 6,301 MKeys
• New middle section activated: iHgzCdB

This means your script will rotate through ~1.54 trillion different middle sections (each taking about 1–8 minutes), while brute-forcing the last 12 characters for each one.  

Have you try With Non Standard Decoding Like a, heuristic Methode  - checksum-ignored Base58 decoding.
it can be faster.

https://drive.google.com/file/d/1ehe1cnNDna8_Sc8VtbTPyF1QhYRKQt9j/view?usp=drive_link
https://drive.google.com/file/d/1UQEGXmGQL509Zu0g1XpLf1raC5pzeuqX/view?usp=drive_link

Wif Ranges 69 is

KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q e crA1gh************
KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q k W1gt2H************


KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q e+(6 middle wif) + 12 Char

KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q f+(6 middle wif) + 12 Char

KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q g+(6 middle wif) + 12 Char

KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q h+(6 middle wif) + 12 Char

KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q i+(6 middle wif) + 12 Char

KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q j+(6 middle wif) + 12 Char

KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3q k+(6 middle wif) + 12 Char


and the most important we need GPU Version, i am not familiar with Cuda and C++ Programming