Gonka community portal, created and maintained by community members

Gonka published a DevShard v6 prerelease and clarified how to submit vulnerability reports and receive payment.

LLM tokens this week
784.9B
DevShard · input + output
GNK this week
-2.2%
Pool trading volume
$349k
Uniswap v3 WGNK/USDT 0.3% on Ethereum
Reporting period
31—6
October 31 — November 6, 2026, UTC
LLM tokens per day
LLM tokens per day065.4B130.9B31 Oct6 Nov
Current periodPrevious period
DevShard · input + output
GNK price · daily close, USD
GNK price · daily close, USD$0.140$0.158$0.17631 Oct6 Nov

DevShard v6 and private reports

A DevShard v6.0.0 prerelease was published on November 2, and the high-availability installation guide was updated. Upgrades without service interruption are supported only for protocol v6 and only when no database restart is required. Local replicas protect against container failure, but not failure of the entire machine. Publication of the build does not confirm its deployment on the mainnet.

On November 5, the documentation clarified that vulnerability reports must first be submitted through HackerOne. An additional private copy on GitHub is not paid for. A report may qualify for a reward after validation, even before a fix is completed; payment for the fix itself is discussed separately and is not guaranteed. The program remains private for now.

GNK fell 2.25%

GNK fell from $0.1790 to $0.1749 over the week in the pool reviewed, a decline of 2.25%. Trading volume totaled $349.4 thousand, compared with $293.0 thousand in the previous full week. These prices do not represent all markets for the token.

Fewer completed requests

DevShard statistics showed 20.99 million completed requests, down 14.19% from the previous week. It processed 766.16 billion input tokens and 18.70 billion output tokens. DeepSeek-V4-Flash-0731 and GLM-5.3-Flash led by total token count, with shares of 34.41% and 34.03%.

Compute capacity metrics and integration proposals

Participants debated the reliability of the dashboard's GPU count metric. Suggestions included removing it from the homepage, switching to an equivalent count of accelerators from another generation, or displaying daily token output. In a related discussion, some participants preferred to release the new mechanism for evaluating computational work before adding models.

An external team was advised to first explain how its technology could apply to the protocol, gather feedback, and prepare a grant proposal. An AMA was discussed as an accompanying future step, with no date set; neither the event nor funding has been confirmed.

← All editions