Ethereum’s Evm.fast Challenge Starts With an Infinite K Bound

The Ethereum Foundation’s evm.fast challenge seeks a machine-checked upper bound, called K, for stateless Ethereum block validation. Its baseline remains infinity, and no finite result has yet been established.

CryptoMEV Bot NewsroomNov 3, 20263 min read

Close-up of multicolored computer code on a dark screen with a shallow depth of field. — Image: One Idea LLC / stocksnap via Openverse (CC0)

At a glance	Details
Who	Ethereum Foundation
What	evm.fast public challenge
When	Nov. 3, 2026
Current result	K is infinity
Input limits	Gas ≤ 200 million; code ≤ 128 MiB

As of Nov (TokenPost). 3, 2026, the Ethereum Foundation had introduced evm.fast, a public challenge inviting contributors to certify an upper bound on the steps required to validate a stateless Ethereum block within a fixed software model. The challenge denotes this ceiling as K. Its baseline is infinity because no finite limit has yet been proved across the covered inputs.

Why is evm.fast measuring K?

According to Crypto Briefing, K certifies an upper limit on execution steps for a 64-bit RISC-V model and must apply to every input in scope, with startup steps included. Flapjack is a Lean 4 adaptation of the Pancake compiler for that architecture; Lean 4 checks mathematical proofs line by line.

The baseline program proves that block verification eventually stops, but it does not establish one fixed ceiling that applies across all inputs. A proven K would give provers a mathematical basis for budgeting the hardest covered block instead of relying on an empirical estimate.

There is also a denial-of-service consideration: a block requiring far more verification steps than expected could leave provers behind, while a certified ceiling would document the defined worst case.

What must entrants submit?

Pluang describes K as the evaluator’s workload across eligible inputs, including startup and execution steps, under limits on gas and program size. It also says the fixed software environment does not guarantee full implementation of Ethereum consensus rules or provide hardware performance metrics.

Eligibility requires an input’s declared gas limit to be 200 million or less, while the submitted program must remain within a 128 MiB program window. The threshold only determines the inputs evaluated by the challenge; Ethereum’s gas rules stay as they are.

A submission must combine RISC-V code with a Lean proof demonstrating that the code matches the fixed Pancake guest. Passing execution tests alone is insufficient.

K is measured in Flapjack’s 64-bit RISC-V model. It reflects evaluator steps rather than Spike’s hardware instruction count, the time needed to produce a proof, or a hardware benchmark.

Why is the baseline still infinity?

At 1:06 a.m. ET on Nov. 3, the baseline was infinity and there were no promoted submissions. Infinity means that no single finite K had been established across all covered inputs.

A kernel replay had been accepted. End-to-end verification of the initial submission, however, remained incomplete.

Both certificate regeneration and the full audit were still in progress.

No monetary prizes, deadlines or submission platforms had been announced for the challenge.

Do Ethereum’s gas rules or on-chain trading conditions change?

Execution proofs are becoming an expected part of Ethereum and are ultimately intended to become mandatory, rather than remaining optional. Such proofs are meant to allow validators to confirm block execution without running it again.

The published 200 million gas eligibility cutoff in evm.fast leaves Ethereum’s gas rules unchanged for readers who use the Ethereum Arbitrage Bot, Aave Flash Loan Arbitrage Bot or Advanced Front-Running Bot for MEV. It defines which inputs the proof-bound challenge evaluates.

A finite result would be valid only for the challenge’s fixed guest and Flapjack RISC-V model. Nor would it show that the guest completely implements Ethereum’s consensus rules, or that the model lines up with a specific zkVM or production platform.

How does evm.fast differ from the Sepolia proof test?

In a separate Sepolia test focused on proof-generation turnaround, one block took 6 to 7 seconds; evm.fast has a different goal: establishing a step ceiling that covers all eligible inputs.

Reported latency for real-time proofs of full Ethereum blocks fell from roughly 16 minutes to under 10 seconds for 99% of mainnet blocks on the target hardware. Real-time proving still depends on budgeting for the hardest possible block rather than only a typical one, which is the specific gap the certified K is intended to close.