[nvpohanh]

[by Claude Code]

Motivation

The cookbook's Agentic Long-Context Deployment sections (#42726, #43083, #43799, #43983, #44065, #44228, #44654, #44712, #45039, #45111, #45311, #45513, #45749, #46005, #46506) mirror SemiAnalysis InferenceX AgentX submissions as raw launch commands, each linked to the run that measured it. The configurations are mirrored from SemiAnalysis submissions. This PR:

1. Adds the MiniMax-M3.1-Flash B200 and GB200 HiCache points at c40, c48 and c56, measured by the newer /use runs of InferenceX #4490 and InferenceX #4491.
2. Syncs to InferenceX main (65a6dd19).
3. Overlays eight open InferenceX PRs (table below).

Source

InferenceX main (65a6dd19), with these open InferenceX PRs overlaid:

PR	Page	Lanes	Pinned to
#4524 MiniMax-M3.1-Flash B300 SGLang TP2	MiniMax-M3.1	minimaxm31flash-fp4-b300-sglang-agentic-dspark-tp2	no /use yet: in progress at head 8eddada3 (same commands as at 2e12ce08 in #46506)
#4491 MiniMax-M3.1-Flash GB200 SGLang	MiniMax-M3.1	minimaxm31flash-fp4-gb200-sglang-agentic-dspark	run 40252912814, named by the PR's /use comment (commit 2751547f; #46506 showed run 40176938051 at d46a2c4f), verified
#4490 MiniMax-M3.1-Flash B200 SGLang	MiniMax-M3.1	minimaxm31flash-fp4-b200-sglang-agentic-dspark	run 40252316011, named by the PR's /use comment (commit 122762e5; #46506 showed run 40176935506 at 96b2c683), verified
#4455 Qwen3.5 FP8 H200 SGLang	Qwen3.5	qwen3.5-fp8-h200-sglang-agentic-mtp, -hicache-mtp	no /use yet: in progress at head fad84b25 (unchanged)
#4434 DeepSeek-V4.1-Flash GB300 Dynamo+SGLang agg	DeepSeek-V4.1	dsv41flash-fp4-gb300-dynamo-sglang-agentic-agg	run 40153045152, named by the PR's /use comment (head 6e0aec9f, same commit as in #46506), verified
#4417 Muse Spark Open B300 SGLang	Muse Spark Open	musespark-fp4-b300-sglang-agentic-dflash	no /use yet: in progress at head 5422503f (unchanged; the PR got new commits right after this PR's sources were frozen, left for the next refresh)
#4346 DeepSeek-V4 NVFP4 B300 Dynamo+SGLang, Mooncake EFA KV transfer	DeepSeek-V4	dsv4-fp4-b300-dynamo-sglang-agentic-agg, -disagg	run 39910889415, named by the PR's /use comment (head 34d4cd6b, unchanged), verified
#4333 Qwen3.5 NVFP4 GB300 Dynamo+SGLang	Qwen3.5	qwen3.5-fp4-gb300-dynamo-sglang-agentic-agg, -disagg; retires -agg-pareto, -pp-pareto	run 39355392627, named by the PR's /use comment (head 02b72dc0, unchanged), verified

"In progress" lanes are rendered at the PR head and labeled "In progress in SA InferenceX AgentX". A later sync pins them once the PR gets a /use run or merges.

• Merged since [Doc] cookbook: Add MiniMax-M3.1-Flash AgentX launches and refresh with nine InferenceX PRs #46506. InferenceX #4492 (MiniMax-M3.1-Flash B300 SGLang) merged and was ingested, so it now comes from main, at the same run (40178345296) and commands as before.
• Retired keys. InferenceX #4333 deletes qwen3.5-fp4-gb300-dynamo-sglang-agentic-agg-pareto and -pp-pareto, as in [Doc] cookbook: Add MiniMax-M3.1-Flash AgentX launches and refresh with nine InferenceX PRs #46506.
• Not mirrored (unmeasured, unchanged from [Doc] cookbook: Add MiniMax-M3.1-Flash AgentX launches and refresh with nine InferenceX PRs #46506): musespark-fp4-h100-sglang-agentic-dflash and qwen3.5-fp8-h100-sglang-agentic. No dashboard changelog names them.
• InferenceX main was frozen for this PR at 65a6dd19. Later merges are left for the next refresh.

What this PR does

Page	Lanes	Deployments	Points	Hardware
MiniMax-M3.1	5	14	52 (was 46)	GB300, GB200, B300, B200
GLM-5.2	14	34	75	GB300, GB200, B300, B200, H200
GLM-5.3	1	1	10	B300
DeepSeek-V4	12	26	70	GB300, GB200, B300, B200, H200
DeepSeek-V4.1	9	22	67	GB300, GB200, B300, B200, H200, H100
Qwen3.5	16	51	220	GB300, GB200, B300, B200, H200, H100
Qwen3.8-Flash-Next	4	5	24	GB300, B300, B200, H200
Muse Spark Open	6	9	64	GB300, GB200, B300, B200, H200

• MiniMax-M3.1-Flash B200 (InferenceX #4490) and GB200 (InferenceX #4491). The TP4 HiCache deployment now covers c32, c40, c48, c56 and c64. The new points use the same serving flags as c32 / c64, with --max-running-requests and --cuda-graph-max-bs-decode at twice the concurrency (80, 96, 112). The other deployments of both lanes keep their commands and now link the newer runs.
• DeepSeek-V4.1-Flash GB300 Dynamo+SGLang agg (InferenceX #4434). Now verified by its /use run; the commands are unchanged.
• Every other page changes only in its source header (InferenceX main SHA and overlay pins). No flag rules were added.

Validation

• Every page in a headless browser (mint dev, Mintlify as in CI). Every hardware / checkpoint / deployment / concurrency / KV tier / router combination renders exactly the decoded commands: 3,138 command sets, 0 mismatches, 0 unseen, no browser errors. Every tab was clicked, and every "Copy link" URL was reopened. The section links from the Deployment panels and the copy buttons were also checked.
• The component's own decoder reproduces every generated plan (3,138 router × KV-tier variants). Every KV flag exists in its image's SGLang commit (239 image × flag pairs). The round-trip / leak check reports 0 problems.
• Arg-parse in the submission images: all 24 commands that changed since the last validated round parse with the lmsysorg/sglang:dev-minimax-m31-flash image's own SGLang parser.
• node scripts/check_cookbook_configs.mjs, mint broken-links --check-anchors --check-redirects, mint validate and pre-commit run --all-files (in lmsysorg/sglang:dev-cu13) pass.

🤖 Generated with Claude Code

---

CI States

Latest PR Test (Base): ✅ Run #40290820249
Latest PR Test (Extra): ❌ Run #40290820190
Latest PR Test (AMD ROCm 10): ➖ No AMD PR run found for this commit.

[mintlify]

Preview approval required

This pull request comes from a fork, so Mintlify won't build a preview until someone with write access approves commit cfe2bfa.

Approve preview