Reported News

Kimi K3.1 Pro live on Fireworks: 2.78T MoE, 64.7% on Sierra τ³-Banking

Fireworks announced on November 4, 2026 that Moonshot's Kimi K3.1 Pro is live on its platform: 2.78T total / 104B active parameters, 1040k input context, and first of 19 on Sierra's τ³-Banking at 64.7% in Fireworks' Specialized Intelligence Index. Serverless pricing is $2.00 uncached input, $0.20 cached input and $10.00 output per million tokens.

Share on X
Changes reported
5
listed below, each from a source we read
We tested
Not tested yet
reported from sources; we did not run it ourselves
Gaps we flag
3
things we could not verify
Sources
1
1 primary, linked below

What changed

Changes the sources report, with the evidence status of each
#What the source saysStatus
01Hosting: Moonshot's existing Kimi K3.1 Pro model is available on Fireworks serverless (pay-per-token) and dedicated deployments; model id accounts/fireworks/models/kimi-k3p1-pro.Reported
02Fireworks restates the model as text and vision, 2.78T total / 104B active parameters, 1040k input context, 896 experts with 16 active per token, using Kimi Delta Attention plus Gated DeepSeek Sparse Attention.Reported
03On Fireworks' Specialized Intelligence Index, K3.1 Pro ranks first of 19 on Sierra's τ³-Banking at 64.7%, 8.7 points ahead of Claude Opus 5.5.Reported
04Other Fireworks-reported SII scores include 61.0% on Factory Legacy Bench (third of 14, top open model) and a tie with Claude Opus 5.5 at 48.2% on Rogo Big Finance Benchmark.Reported
05Standard serverless pricing per million tokens: $2.00 uncached input, $0.20 cached input, $10.00 output.Reported

reported by the sourcemeasured or run by usnot verified

Primary source: Fireworks blog: Kimi K3.1 Pro on Fireworks. We are the fast, independent index; the vendor’s own page is the source of record for this release.

In the source’s words

“It ranks first of 19 on Sierra's τ³-Banking in our Specialized Intelligence Index, 8.7 points ahead of Claude Opus 5.5.”

“K3.1 Pro ranks first of 19 at 64.7%, 8.7 percentage points ahead of Claude Opus 5.5.”

“Standard serverless pricing per million tokens: $2.00 uncached input, $0.20 cached input and $10.00 output.”

Early verdict

A hosting add for an existing Moonshot model, useful if you want Fireworks serverless or dedicated capacity and care about the reported τ³-Banking lead.

Basis: Based on Fireworks' November 4 post; the model itself was already catalogued from Moonshot's earlier launch.

What we have not verified

  • We did not call the Fireworks endpoint or reproduce the Specialized Intelligence Index scores.
  • Parameter and context figures are restated by Fireworks from Moonshot; we did not re-verify them against Moonshot's pages for this hosting brief.
  • Dedicated-deployment pricing is not listed in the post.

This brief is Reported: written from the sources below, not from running the release ourselves.

Sources

Spotted an error? See our corrections policy or email the address on the contact page. Markdown version for agents: /news/kimi-k3-1-pro-on-fireworks.md

Want these briefs in your Google Top Stories? Add Agentic Weekly as a preferred source on Google.

Newsletter

The Weekly Verdict. Free, every Monday.

A free email every Monday: the week's few big AI releases, then 5 short use cases for the new models, each with a prompt you can paste.

Free. Double opt-in. One click to unsubscribe.