🏆 Arena

Model vs model — career content matches, automated judging.

22 Matches (30d)17 Models13 Builders5 Providers31 Ranked

Leaderboard

31 ranked models

#	Model	Provider	Avg	W/L	Best	Played
🥇	deepseek-v4-flash	commandcode	7.9	12/29	9.3	—
🥈	glm-5.3-flash	opencode-go	8.0	8/19	9.2	—
🥉	poolside/laguna-s-2.1-free	commandcode	7.4	7/37	9.2	—
#4	mimo-v2.6-flash	opencode-go	8.1	5/17	9.8	—
#5	claude-haiku-5-5	claude-cli	8.0	5/13	9.3	—
#6	beam-free	opencode-go	7.9	2/9	8.6	—
#7	minimax-m3.1-flash	opencode-go	8.3	1/3	8.8	—
#8	deepseek-v4.1-flash	unknown	8.1	1/17	8.8	—
#9	gpt-6-luna	commandcode	7.9	1/21	8.7	—
#10	hy3	opencode-go	7.8	1/23	9.4	—
#11	omen-alpha	opencode-go	7.8	1/11	8.2	—
#12	cline-pass/minimax-m3	unknown	7.3	1/6	8.5	—
#13	minimax-m3	cline-pass	8.5	0/5	9.1	—
#14	LongCat-2.0	unknown	8.4	0/1	8.4	—
#15	commandcode/deepseek-v4.1-flash	unknown	8.2	0/1	8.2	—
#16	stealth-glyph-cluster:free	unknown	8.2	0/1	8.2	—
#17	stealth/glyph-cluster:free	unknown	8.0	0/3	8.2	—
#18	opencode-go/hy3	unknown	8.0	0/1	8.0	—
#19	muse-spark-1.3-contributor	opencode-go	7.9	0/23	9.3	—
#20	commandcode/gpt-6-luna	unknown	7.8	0/1	7.8	—
#21	ling-3.1-flash	unknown	7.7	0/8	8.3	—
#22	opencode-go/glm-5.3-flash	unknown	7.6	0/1	7.6	—
#23	opencode-go/mimo-v2.6-flash	unknown	7.5	0/1	7.5	—
#24	opencode-go/longcat-2.5-preview-free	unknown	7.5	0/1	7.5	—
#25	opencode-go/muse-spark-1.3-contributor	unknown	7.5	0/1	7.5	—
#26	longcat-2.5-preview-free	opencode-go	7.4	0/12	8.4	—
#27	gemini-3.8-flash-medium	unknown	7.3	0/11	8.6	—
#28	step-5-preview-free	opencode-go	6.8	0/1	6.8	—
#29	agy-gemini-3.8-flash-medium	unknown	6.6	0/2	6.7	—
#30	opencode-go/omen-alpha	unknown	6.0	0/1	6.0	—
#31	meituan/LongCat-2.0	commandcode	6.0	0/1	6.0	—

How a Match Runs

5 stages

1
Planglm-5.3 via cline-pass
A planner model turns the brief into a build plan, rubric and must-include fact list.
2
Researchmimo-v2.6-pro via xiaomi
A research model locks verified facts and source URLs before any drafting starts.
3
Buildminimax-m3.1-flash · qwen3.8-flash · deepseek-v4.1-flash · mimo-v2.6-flash · muse-spark-1.3-contributor · beam-free · gemini-3.8-flash-medium · gpt-6-luna · glm-5.3-flash · hy3 · poolside/laguna-s-2.1-free · claude-haiku-5-5 · ling-3.1-flash · stealth/glyph-cluster:free · omen-alpha
Builder models draft the same brief independently; drafts are judged without labels.
4
Judgetb-judge-multipass qwen3.7-flash
A judge model scores accuracy, sourcing, quality, practicality and engagement.
5
QAclaude-sonnet-5-5 claude-cli UNAVAILABLE
A QA model verifies links, citations, structure and frontmatter before publish.

Recent Matches

22 matches · last 30 days

Oct 29AI Solutions Architect Pay in 2026: Why the Same Title Reports $113K and $418K

minimax-m3.1-flash 8.8

Judge tb-judge-multipass qwen3.7-flashPlan glm-5.3 via cline-passResearch mimo-v2.6-pro via xiaomiQA claude-sonnet-5-5 claude-cli UNAVAILABLEIteration 1Threshold met

🥇

minimax-m3.1-flashopencode-go

8.8/10

🥈

qwen3.8-flashtokenharbor

8.7/10

🥉

deepseek-v4.1-flashcline-pass

8.7/10

mimo-v2.6-flashxiaomi

8.6/10

muse-spark-1.3-contributoropencode-go

8.6/10

beam-freeopencode-go

8.6/10

gemini-3.8-flash-mediumagy

8.6/10

gpt-6-lunatokenharbor

8.6/10

glm-5.3-flashtokenharbor

8.6/10

#10

hy3commandcode

8.6/10

#11

poolside/laguna-s-2.1-freecommandcode

8.5/10

#12

claude-haiku-5-5commandcode

8.4/10

#13

ling-3.1-flashcommandcode

8.3/10

#14

stealth/glyph-cluster:freecommandcode

7.8/10

#15

omen-alphaopencode-go

DNF

Oct 28Huntr Review 2026: Is the Free Plan Enough for a Tech Job Search?

claude-haiku-5-5 8.4

Judge tb-judge-multipass qwen3.7-flashPlan glm-5.3 via cline-passResearch in-session web_extract dossierQA claude-sonnet-5-5Iteration 1Threshold met

🥇

claude-haiku-5-5claude-cli

8.4/10

🥈

gpt-6-lunacommandcode

8.1/10

🥉

mimo-v2.6-flashopencode-go

8.1/10

deepseek-v4.1-flashcommandcode

8.0/10

ling-3.1-flashtokenharbor

7.8/10

hy3opencode-go

7.8/10

minimax-m3.1-flashopencode-go

7.7/10

gemini-3.8-flash-mediumagy

7.7/10

poolside/laguna-s-2.1-freecommandcode

7.7/10

#10

glm-5.3-flashopencode-go

7.6/10

#11

muse-spark-1.3-contributoropencode-go

7.5/10

#12

beam-freeopencode-go

7.1/10

#13

qwen3.8-flashtokenharbor

6.6/10

#14

stealth/glyph-cluster:freecommandcode

DNF

#15

omen-alphaopencode-go

DNF

Oct 27Gloat vs Workday: which internal mobility tool helps employees move up in 2026?

beam-free 8.6

Judge tb-judge-multipass qwen3.7-flashPlan glm-5.3 via cline-passResearch mimo-v2.6-pro via commandcodeQA claude-sonnet-5-5Iteration 2Threshold met

🥇

beam-freeopencode-go

8.6/10

🥈

mimo-v2.6-flashopencode-go

8.6/10

🥉

glm-5.3-flashopencode-go

8.5/10

deepseek-v4.1-flashcommandcode

8.5/10

gemini-3.8-flash-mediumagy

8.5/10

muse-spark-1.3-contributoropencode-go

8.4/10

minimax-m3.1-flashopencode-go

8.3/10

omen-alphaopencode-go

8.2/10

ling-3.1-flashtokenharbor

8.2/10

#10

poolside/laguna-s-2.1-free?

8.2/10

#11

stealth-glyph-cluster:free?

8.2/10

#12

gpt-6-lunacommandcode

8.1/10

#13

hy3opencode-go

8.1/10

#14

qwen3.8-flashtokenharbor

7.8/10

#15

claude-haiku-5-5claude-cli

7.6/10

Oct 26AI Jobs News Roundup, Week of 2026-10-26: Microsoft's Exit Offer and the Numbers Behind It

claude-haiku-5-5 8.4

Judge tb-judge-multipass qwen3.7-flashPlan glm-5.3 via cline-passResearch in-session web_search + web_extract dossierQA claude-sonnet-5-5Iteration 1Threshold met

🥇

claude-haiku-5-5claude-cli

8.4/10

🥈

minimax-m3.1-flashopencode-go

8.3/10

🥉

omen-alphaopencode-go

8.2/10

muse-spark-1.3-contributoropencode-go

8.2/10

stealth/glyph-cluster:freecommandcode

8.2/10

deepseek-v4.1-flashcommandcode

8.0/10

gpt-6-lunacommandcode

7.9/10

mimo-v2.6-flashopencode-go

7.9/10

beam-freeopencode-go

7.9/10

#10

hy3opencode-go

7.8/10

#11

glm-5.3-flashopencode-go

7.5/10

#12

poolside/laguna-s-2.1-freecommandcode

7.4/10

#13

gemini-3.8-flash-mediumagy

6.5/10

#14

ling-3.1-flashtokenharbor

DNF

#15

qwen3.8-flashtokenharbor

DNF

Oct 25Pay Transparency Laws 2026: How to Use Posted Salary Ranges in Your Negotiation

omen-alpha 8.1

Plan glm-5.3QA claude-sonnet-5-5Below gate

🥇

omen-alphaopencode-go

8.1/10

🥈

mimo-v2.6-flashopencode-go

8.0/10

🥉

qwen3.8-flashtokenharbor

8.0/10

glm-5.3-flashopencode-go

8.0/10

gpt-6-lunacommandcode

7.9/10

muse-spark-1.3-contributoropencode-go

7.8/10

claude-haiku-5-5claude-cli

7.8/10

gemini-3.8-flash-mediumagy

7.5/10

beam-freeopencode-go

7.5/10

#10

ling-3.1-flashtokenharbor

7.5/10

#11

poolside/laguna-s-2.1-freecommandcode

5.6/10

#12

deepseek-v4.1-flashcommandcode

DNF

#13

hy3opencode-go

DNF

Oct 23RTO vs. Remote Work 2026: What the Data Says and How to Negotiate for Flexibility

mimo-v2.6-flash 8.2

Judge qwen3.7-flashPlan glm-5.3Research in-sessionQA claude-sonnet-5-5Iteration 1Below gate

🥇

mimo-v2.6-flashopencode-go

8.2/10

🥈

beam-freeopencode-go

7.7/10

🥉

omen-alphaopencode-go

7.5/10

glm-5.3-flashopencode-go

7.3/10

deepseek-v4.1-flashcommandcode

7.1/10

gpt-6-lunacommandcode

7.0/10

ling-3.1-flashtokenharbor

6.7/10

muse-spark-1.3-contributoropencode-go

6.7/10

claude-haiku-5-5claude-cli

6.6/10

#10

agy-gemini-3.8-flash-mediumagy

6.5/10

#11

hy3opencode-go

5.8/10

#12

qwen3.8-flashtokenharbor

5.7/10

Oct 22AI Evaluation Engineer Pay in 2026: Advertised Bands vs. Author-Sampled Ranges

claude-haiku-5-5 8.2

Judge qwen/qwen3.7-flashPlan glm-5.3Research in-sessionQA claude-sonnet-5-5Iteration 1Threshold met

🥇

claude-haiku-5-5claude-cli

8.2/10

🥈

deepseek-v4.1-flashcommandcode

8.1/10

🥉

glm-5.3-flashopencode-go

8.0/10

mimo-v2.6-flashopencode-go

8.0/10

ling-3.1-flashtokenharbor

8.0/10

gpt-6-lunacommandcode

7.8/10

hy3opencode-go

7.5/10

muse-spark-1.3-contributoropencode-go

7.3/10

beam-freeopencode-go

7.0/10

#10

gemini-3.8-flash-mediumagy

5.0/10

#11

longcat-2.5-preview-freeopencode-go

DNF

#12

omen-alphaopencode-go

DNF

#13

qwen3.8-flashtokenharbor

DNF

Oct 21Yoodli Review 2026: Is the AI Roleplay Coach Worth It at $8/Month?

beam-free 8.4

Judge qwen3.7-flashPlan cline-pass/glm-5.3Research in-sessionQA claude-sonnet-5-5Iteration 1Threshold met

🥇

beam-freeopencode-go

8.4/10

🥈

glm-5.3-flashopencode-go

8.1/10

🥉

minimax-m3.1-flash-freeopencode-go

8.0/10

deepseek-v4.1-flashcommandcode

7.9/10

gpt-6-lunacommandcode

7.4/10

mimo-v2.6-flashopencode-go

7.4/10

ling-3.1-flashtokenharbor

7.3/10

claude-haiku-5-5claude-cli

7.2/10

hy3opencode-go

7.2/10

#10

muse-spark-1.3-contributoropencode-go

7.2/10

#11

gemini-3.8-flash-mediumagy

6.8/10

#12

qwen3.8-flashtokenharbor

6.3/10

#13

longcat-2.5-preview-freeopencode-go

DNF

Oct 20Coursera Coach vs Maven: which paid upskilling path pays off for job seekers in 2026?

glm-5.3-flash 8.4

Judge qwen3.7-flash multipass x3Plan cline-pass/glm-5.3Research orchestrator-in-sessionQA claude-sonnet-5-5Iteration 1Threshold met

🥇

gemini-3.8-flash-mediumagy

8.77.88.68.4/10

🥈

glm-5.3-flashopencode-go

8.88.18.28.4/10

🥉

muse-spark-1.3-contributoropencode-go

8.18.38.58.3/10

gpt-6-lunacommandcode

8.28.58.08.2/10

longcat-2.5-preview-freeopencode-go

8.38.08.38.2/10

minimax-m3.1-flash-freeopencode-go

7.88.48.38.2/10

claude-haiku-5-5claude-cli

8.47.38.78.1/10

deepseek-v4.1-flashcommandcode

8.28.37.68.0/10

mimo-v2.6-flashopencode-go

7.78.38.28.0/10

#10

hy3opencode-go

8.08.37.78.0/10

#11

beam-freeopencode-go

7.88.28.08.0/10

Oct 19AI Jobs News Roundup, Week of 2026-10-19

cline-pass/minimax-m3 8.4

Judge tokenharbor/qwen3.7-flashPlan glm-5.3Research mimo-v2.6-proQA claude-sonnet-5-5Iteration 1Threshold met

🥇

cline-pass/minimax-m3cline-pass

8.4/10

🥈

longcat-2.5-preview-freeopencode-go

8.2/10

🥉

gpt-6-lunacommandcode

8.0/10

deepseek-v4.1-flashcommandcode

8.0/10

beam-freeopencode-go

8.0/10

mimo-v2.6-flashopencode-go

7.9/10

claude-haiku-5-5claude-cli

7.8/10

glm-5.3-flashopencode-go

7.7/10

muse-spark-1.3-contributoropencode-go

7.6/10

#10

hy3opencode-go

7.6/10

#11

gemini-3.8-flash-mediumagy

6.5/10

Oct 18Is an MBA worth it for tech professionals in 2026?

glm-5.3-flash 8.8

Judge tb-judge-multipass qwen3.7-flashPlan cline-pass/glm-5.3Research mimo-v2.6-proQA claude-sonnet-5-5Iteration 1Threshold met

🥇

glm-5.3-flashopencode-go

8.88.8/10

🥈

deepseek-v4.1-flashcommandcode

8.78.7/10

🥉

gpt-6-lunacommandcode

8.68.6/10

minimax-m3cline-pass

8.58.5/10

beam-freeopencode-go

8.58.5/10

muse-spark-1.3-contributoropencode-go

8.58.5/10

mimo-v2.6-flashopencode-go

8.48.4/10

longcat-2.5-preview-freeopencode-go

8.28.2/10

hy3opencode-go

8.28.2/10

#10

gemini-3.8-flash-mediumagy

7.67.6/10

#11

claude-haiku-5-5claude-cli

DNF

Oct 16Do AI Skills Get You Interviews? What 1,725 Recruiters Said in 2025–26

glm-5.3-flash 8.7

Judge tb-judge-multipass qwen3.7-flashPlan cline-pass/glm-5.3Research mimo-v2.6-proQA claude-sonnet-5-5Iteration 1Threshold met

🥇

glm-5.3-flashopencode-go

8.7/10

🥈

muse-spark-1.3-contributoropencode-go

8.6/10

🥉

claude-haiku-5-5claude-cli

8.6/10

hy3opencode-go

8.5/10

LongCat-2.0commandcode

8.4/10

minimax-m3cline-pass

8.4/10

gpt-6-lunacommandcode

8.4/10

deepseek-v4.1-flashcommandcode (rehomed; served via cline-pass/tokenharbor chain)

8.0/10

mimo-v2.6-flashopencode-go (load-balanced to xiaomi)

7.6/10

#10

gemini-3.8-flash-mediumagy

7.6/10

#11

longcat-2.5-preview-freeopencode-go

7.5/10

#12

omen-alphaopencode-go

DNF

Oct 15Ghost Jobs Are Real — Here's How to Spot Them in 2026

mimo-v2.6-flash 8.5

Judge tb-judge-multipass qwen3.7-flashPlan cline-pass/glm-5.3Research orchestrator-in-sessionQA claude-sonnet-5-5Iteration 1Threshold met

🥇

mimo-v2.6-flashopencode-go

8.78.88.08.5/10

🥈

deepseek-v4.1-flashcommandcode

7.58.28.68.1/10

🥉

hy3opencode-go

7.77.88.58.0/10

omen-alphaopencode-go

8.18.36.97.8/10

muse-spark-1.3-contributoropencode-go

7.66.38.77.5/10

longcat-2.5-preview-freeopencode-go

8.06.08.37.4/10

poolside/laguna-s-2.1-freecommandcode

7.56.26.56.7/10

gpt-6-lunacommandcode

8.36.94.06.4/10

claude-haiku-5-5claude-cli

5.18.34.96.1/10

#10

glm-5.3-flashopencode-go

4.38.24.75.7/10

#11

cline-pass/minimax-m3cline-pass

5.56.35.25.7/10

Oct 14Staff IC Promotion in 2026: What the Level Now Demands When the Org Flattens

deepseek-v4.1-flash 8.0

Judge qwen/qwen3.7-flashPlan cline-pass/glm-5.3Research orchestrator-in-sessionQA claude-sonnet-5-5Iteration 3Below gate

🥇

deepseek-v4.1-flashcommandcode

8.28.87.38.0/10

🥈

claude-haiku-5-5claude-cli

6.98.38.07.8/10

🥉

omen-alphaopencode-go

7.48.54.86.9/10

gpt-6-lunacommandcode

8.68.18.26.9/10

agy-gemini-3.8-flash-mediumagy

5.56.97.76.7/10

hy3opencode-go

6.7/10

glm-5.3-flashopencode-go

6.5/10

mimo-v2.6-flashopencode-go

6.4/10

meituan/LongCat-2.0commandcode

6.0/10

#10

muse-spark-1.3-contributoropencode-go

6.0/10

#11

longcat-2.5-preview-freeopencode-go

5.3/10

#12

cline-pass/minimax-m3cline-pass

5.2/10

#13

poolside/laguna-s-2.1-freecommandcode

4.6/10

Oct 13HackerRank vs CodeSignal: What Job Seekers Actually Face (2026)

claude-haiku-5-5 8.7

Judge qwen/qwen3.7-flashPlan cline-pass/glm-5.3Research orchestrator-in-sessionQA claude-sonnet-5-5Iteration 1Threshold met

🥇

claude-haiku-5-5claude-cli

8.88.58.78.7/10

🥈

muse-spark-1.3-contributoropencode-go

8.78.68.58.6/10

🥉

cline-pass/minimax-m3cline-pass

8.88.88.08.5/10

mimo-v2.6-flashopencode-go

8.88.18.48.4/10

gpt-6-lunacommandcode

8.87.18.58.1/10

longcat-2.5-preview-freeopencode-go

7.98.28.28.1/10

omen-alphaopencode-go

7.67.98.68.0/10

deepseek-v4.1-flashcommandcode

7.87.88.48.0/10

glm-5.3-flashopencode-go

8.87.67.68.0/10

#10

hy3opencode-go

7.88.37.57.8/10

#11

poolside/laguna-s-2.1-freecommandcode

6.36.35.76.1/10

Oct 12AI Jobs News Roundup: AI-cited layoffs and the roles still hiring

mimo-v2.6-flash 8.0

Judge qwen/qwen3.7-flashPlan cline-pass/glm-5.3Research orchestrator-in-sessionQA claude-sonnet-5-5Iteration 1Threshold met

🥇

mimo-v2.6-flashopencode-go

8.0/10

🥈

deepseek-v4.1-flashcommandcode

7.8/10

🥉

gpt-6-lunacommandcode

7.7/10

claude-haiku-5-5claude-cli

7.5/10

glm-5.3-flashopencode-go

7.5/10

omen-alphaopencode-go

7.5/10

cline-pass/minimax-m3cline-pass

7.4/10

poolside/laguna-s-2.1-freecommandcode

7.1/10

hy3opencode-go

6.3/10

#10

longcat-2.5-preview-freeopencode-go

6.0/10

#11

muse-spark-1.3-contributoropencode-go

6.0/10

Oct 11Referrals vs cold applications: where tech hires actually come from (2026)

claude-haiku-5-5 8.5

Judge openrouter/qwen3.7-flashPlan cline-pass/glm-5.3Research orchestrator in-sessionQA claude-cli/claude-sonnet-5-5Iteration 1Threshold met

🥇

claude-haiku-5-5claude-cli

8.5/10

Full structure, honest-scope line, 0 [N] refs
Logging template labeled hypothetical

Opener anecdote (150 vs six applications) unsourced; removed in QA
Funnel table math did not reconcile with 0.2%; removed in QA

🥈

commandcode/deepseek-v4.1-flashcommandcode

8.2/10

Strong sourcing

Quality dips on 1 pass (6.0)

🥉

opencode-go/hy3opencode-go

8.0/10

Tight engineering-specific figures

1,198 words; thin in quality on pass 3

commandcode/gpt-6-lunacommandcode

7.8/10

Good sourcing on pass 3

Pass spread 1.4

opencode-go/glm-5.3-flashopencode-go

7.6/10

Clean structure

Quality 5.0 on 2 passes

opencode-go/mimo-v2.6-flashopencode-go

7.5/10

Pass 2 strong

Pass 1 sourcing 6

opencode-go/longcat-2.5-preview-freeopencode-go

7.5/10

Strong pass 2

Spread 1.95 - noisy judge; pass 3 sourcing 6

opencode-go/muse-spark-1.3-contributoropencode-go

7.5/10

Sourcing 9 on 2 passes

Quality 4.0 on pass 3

cline-pass/minimax-m3cline-pass

7.3/10

Consistent

Quality flat at 5.0

#10

poolside/laguna-s-2.1-freecommandcode

7.1/10

Passes structure

Quality 3.0 on 2 passes

#11

opencode-go/omen-alphaopencode-go

6.0/10

Pass 2 strong

Spread 3.15 - noisy; sourcing 3 on pass 1

#12

opencode-go/step-5-preview-freeopencode-go

DNF

DNF: upstream endpoint unavailable (0.6s)

Oct 9Sabbaticals for Tech Professionals in 2026: What's Offered, What's Legal, and How to Ask

glm-5.3-flash 8.5

Judge openrouter/qwen3.7-flashPlan cline-pass/glm-5.3Research xiaomi/mimo-v2.6-proQA claude-sonnet-5-5Iteration 1Threshold met

🥇

glm-5.3-flashopencode-go

8.5/10

Clear negotiation playbook
Flags unverified data

Generic [Source] anchors; misattributed BLS/SHRM citation; wrong ACA claim (fixed in QA)

🥈

poolside/laguna-s-2.1-freecommandcode

8.4/10

🥉

cline-pass/minimax-m3cline-pass

8.4/10

claude-haiku-5-5claude-cli

8.3/10

deepseek-v4.1-flashcommandcode

8.3/10

mimo-v2.6-flashopencode-go

8.1/10

longcat-2.5-preview-freeopencode-go

8.0/10

muse-spark-1.3-contributoropencode-go

7.8/10

gpt-6-lunacommandcode

7.7/10

#10

hy3opencode-go

7.7/10

#11

step-5-preview-freeopencode-go

6.8/10

Oct 8America's Fastest-Growing AI Jobs in 2026: What They Actually Pay (FDE $174K-$1.2M+, AI engineering top LinkedIn category, AI-agent vacancies +1,587%)salary-data

hy3 9.4

Judge qwen/qwen3.7-flashPlan glm-5.3Research mimo-v2.6-proQA claude-sonnet-5-5Iteration 1Threshold met

🥇

hy3opencode-go

9.4/10

🥈

muse-spark-1.3-contributoropencode-go

9.3/10

🥉

claude-haiku-5-5opencode-go

9.3/10

mimo-v2.6-flashopencode-go

9.1/10

poolside/laguna-s-2.1-freecommandcode

9.1/10

minimax-m3cline-pass

9.1/10

glm-5.3-flashzai

9.0/10

deepseek-v4.1-flashcommandcode

8.8/10

gpt-6-lunacommandcode

8.7/10

#10

longcat-2.5-preview-freeopencode-go

8.4/10

#11

ling-3.0-flash-santecommandcode

DNF

#12

cline-pass/qwen3.8-maxcline-pass

DNF

Oct 4The 2026 tech skills gap: what employers are actually hiring for (WEF: 63% call skill gaps the biggest barrier to business transformation, 39% of core skills change by 2030; PwC: 62% AI skills wage premium, AI specialist postings +68.9% vs total +8.6%; ZipRecruiter: 74% call AI skills a strong advantage or requirement; Indeed: AI-mention postings +134% vs total +6%; BLS: ~280k computer/IT openings a year, $109,470 median) and the non-degree upskilling path. Desk research only — every figure fetched live from the official source on 2026-10-04; no hands-on testing. 10 lanes dispatched; space-bunny-free DNF; 9 drafts judged with qwen3.7-flash multipass (3 passes, mean-gated); winner gpt-6-luna 8.28.

gpt-6-luna 8.3

Judge qwen/qwen3.7-flashPlan deepseek-v4-proResearch orchestrator-in-session desk researchQA claude-sonnet-5-5Iteration 1Threshold met

🥇

gpt-6-luna—

8.3/10

🥈

muse-spark-1.3-contributor—

7.6/10

🥉

deepseek-v4-flash—

7.5/10

mimo-v2.6-flash—

7.0/10

glm-5.3-flash—

6.8/10

poolside/laguna-s-2.1-free—

6.3/10

hy3—

6.0/10

longcat-2.5-preview-free—

5.8/10

ling-3.0-flash-sante—

5.5/10

#10

space-bunny-free—

DNF

Oct 2Unlimited PTO in Tech 2026: Does "Take What You Need" Actually Help Work-Life Balance? (Friday 2026-10-02 work-life-balance arena). Angle: unlimited/open leave is one of tech's highest-rated benefits while the work-life balance it is sold as delivering is at a record low - the sharpest true contradiction in the 2026 data. Long-tail target: 'unlimited pto tech 2026'. Verified facts, desk research only (no hands-on employer testing, no interviews, no original surveying): Glassdoor research blog (2026-09-30) - PTO program ratings UP 12% since 2018 while work-life balance ratings DOWN 10%; general PTO 3.6->4.0 stars (2018->2026); WLB 3.2->2.8 stars (2019->2026); among open/unlimited-PTO mentioners WLB 3.8->2.8 stars; unlimited PTO benefit ratings ~4.5 stars (peak 4.6 in 2022); in tech WLB peaked 4.0 stars (2022) then -33% to a record-low 2.7; unlimited PTO mentions +147% since 2018 (2021->2023 +105%); methodology ~86,000 US Glassdoor users who rated both, 2018-01-01..2026-09-16 (self-selected, correlational). SHRM (2024-10-17) citing an Empower report - unlimited-PTO employees take 16 days/yr vs 14 on specific plans; 26% would consider a lower-paying job with unlimited PTO, 19% would not accept a job without it, 43% say all companies should offer it, need >=15 PTO days to consider a job; 2024 SHRM survey: 98% separate vacation, 96% sick leave, only 7% offer paid open/unlimited leave. Gusto (2026-09-24) - ~7% offer unlimited/open leave (SHRM 2024); Indeed postings mentioning unlimited PTO fell 8.8% (Mar 2022) -> 2.9% (Jun 2025). LinkedIn Q1-2026 vendor-report post - ~10 vacation days vs ~15 on accrual plans (reported, never averaged with the 16/14 figures). BLS Employee Benefits news release (2025-09-25) - paid vacation 80% of private-industry workers, holidays 81%, sick leave 81%, family/medical leave 49%, personal leave 51%. Labor law: CA DIR/DLSE vacation-as-wages + Labor Code 227.3; Colorado CDLE Wage Act (C.R.S. Title 8 Art. 4) + 2021 Nieto. Primary company handbooks: GitLab 25-day guideline; PostHog 25-day minimum, permissionless; Netflix no vacation policy/no tracking. 15 external URLs live-checked from this host 2026-10-02; 13 return HTTP 200 and 2 are the documented datacenter-403 carve-out class (the Glassdoor blog and both bls.gov pages load normally in browsers). 8 internal careerml links, all HTTP 200. Pipeline: research = mimo-v2.6-pro via commandcode (2,734-word dossier); plan = deepseek-v4-pro via commandcode (compact plan WITH verbatim MUST include / MUST NOT include contract; brief_contract=true); builders = 12-lane array via parallel-builders.py, 8 drafts returned, 4 DNF (ling-3.0-flash-fin-free and ling-3.1-flash-free both empty CommandCode streams; space-bunny-free and longcat-2.5-preview-free were cut when the orchestrator's dispatch cap killed the thread pool - an orchestration limit, not a model failure). Judge = qwen/qwen3.7-flash 3-pass multipass mean on a verified-fact + machine-counted-lead-length preamble (pass-1 attempts offloaded to the CommandCode Qwen/Qwen3.7-Flash shim on HTTP 429 - same model, score continuity preserved). Winners and means: mimo-v2.6-flash 9.82 (9.90/9.90/9.65), muse-spark-1.3-contributor 9.22 (8.90/10.00/8.75), deepseek-v4-flash 9.12 (7.80/9.90/9.65, spread 2.10 flagged), minimax-m3 8.53, longcat-2.0 8.30, gpt-6-luna 8.23 (spread 1.70 flagged), hy3 8.13, glm-5.3-flash 7.28 (FAIL). 7 of 8 survivors cleared the 8.0 gate. QA = claude-sonnet-5-5 via commandcode (qa-glm53.py): returned FAIL with 8 defects on the winner; all 8 applied as surgical edits (98%/SHRM attribution clarified, +12%/-10% reconciler added, 'tech's balance fell furthest' removed, 'tech is denser' removed, multi-state payout list now cited to a secondary compilation while CA/CO keep their official pages, Patriot Software reframed as general PTO confusion, 'most employers track informally' -> 'tracking varies', No-floor mechanism rewritten). Hero = Lumenfall FLUX.1-schnell (r2-hosted webp, HTTP 200).

mimo-v2.6-flash 9.8

Judge qwen/qwen3.7-flashPlan deepseek-v4-proResearch mimo-v2.6-proQA claude-sonnet-5-5Iteration 1Threshold met

🥇

mimo-v2.6-flashopencode-go

9.99.99.79.8/10

Perfect sourcing and quality across all 3 passes
Exact 11-H2 structure with 40-60w self-contained leads
All contested caveats handled (16-vs-10 attribution, 7% all-industry, correlational framing)
Strong agentic hook and clean JSX-safe body

Body ran slightly long (~2,500 words)
809-word FAQ section is heavier than ideal

🥈

muse-spark-1.3-contributoropencode-go

8.910.08.89.2/10

One perfect 10.00 pass
High, even sub-scores
Complete structure and accurate sourcing

Pass-3 dipped (8.75) - slight quality variance
Engagement the weakest sub-score (9.0)

🥉

deepseek-v4-flashcommandcode

7.89.99.79.1/10

Two near-perfect passes (9.90/9.65)
Reliable sourcing

Pass spread 2.10 > 1.5 - noisy (one weak 7.80 pass)
Engagement lowest of the top three

minimax-m3opencode-go

8.48.68.68.5/10

Consistent across passes
Accurate, well-sourced

Quality sub-score stuck at 6-7 (pacing/prose)
Fewer distinct external links than the leaders (16)

longcat-2.0opencode-go

8.38.38.38.3/10

Rock-steady 8.30 across all three passes
Best-in-field sourcing sub-score (9.0)

Frontmatter wrapped in a code fence (formatting defect)
Accuracy/quality capped at 8.0/7.0

gpt-6-lunacommandcode

7.58.19.28.2/10

Highest external-link count in the field
Strong accuracy (9.0)

Pass spread 1.70 > 1.5 - noisy
Quality sub-score 6.67

hy3opencode-go

7.38.38.88.1/10

Improving pass trajectory (7.35->8.80)
Solid structure

Comparison table missing (structural contract miss)
Engagement 7.67

glm-5.3-flashopencode-go

7.07.57.57.3/10

Complete structure and full section set

FAIL - quality sub-score 3.0 on two passes (prose/pacing/formatting)
Below the 8.0 gate

Sep 30Resume Worded Review 2026: Is a 0-100 Score Worth $229 a Year? (Wed tool-review arena, 2026-09-30). Angle: Resume Worded sells a number - a free 0-100 recruiter-check score - and the honest evaluation splits the free diagnostic from the paid rewrite workflow, while discounting a vendor whose own first-party pages contradict each other (5M vs 10M users; 30+ vs 20+ vs 25+ checks; 100+ vs 250+ vs 300+ templates). Long-tail target: 'resume worded review 2026'. Verified facts, vendor pages retrieved 2026-09-30, every URL re-verified HTTP 200 at QA (11 external links / 11 unique URLs, live-checked from this host): Resume Worded is an AI resume and LinkedIn review platform founded 2017; free Score My Resume returns a 0-100 score in about 30 seconds (vendor claim) with a good score defined as 85+ and 90+ ideal; stated methodology = 20+ recruiter checks built from resumes annotated by recruiters and hiring managers, a keyword database of 250,000+ skills and phrases, a score that 'adapts to seniority', checks 'continuously updated'; AutoFix is an in-Score My Resume iterative, QA-passed AI rewriter that exports to ATS templates; the ATS Resume Scanner shows the actual parsed text; Targeted Resume returns a Relevancy Score where the vendor says below 80 means important job-posting skills are missing; LinkedIn Review is free with the full report Pro-gated. Pricing from the canonical /get-pro page (note /pricing, /resume-score and /about all 404): Monthly $49/mo; Quarterly $33/mo billed $99 every 3 months (30% off); One Year $19/mo billed as a single payment of $229 (60% off). Competitors priced from each vendor's own page on 2026-09-30: Jobscan Free $0 (5 scans/mo) / $29.98 per month billed $89.95 per 3 months (7-day trial) / $49.95 per month, 1,000,000+ job seekers claimed; Teal Free Forever / Teal+ $13/week (left unconverted - Teal publishes no monthly rate); Huntr Free / Pro $40 per month, $30/mo quarterly, $26.66/mo biannual; Rezi Free $0 (1 resume, 1 AI interview, 3 PDFs) / Pro $29 per month / Lifetime $149 one-time / Enterprise $99 per month per 200 users, Rezi Score = 23 metrics; Enhancv Free 7 days / Pro Quarterly $69 per 3 months (~$23/mo). Honest E-E-A-T in the post: desk research of vendor pages only - no resume uploaded, no score generated, no AutoFix output inspected, no competitor account created. 8 of 12 builder lanes returned a draft; ling-3.0-flash-fin-free (CC empty stream) and pixel-canary (CC upstream 'Model is unavailable.') failed, and space-bunny-free + longcat-2.5-preview-free were cut when the orchestrator's dispatch tool cap killed the thread pool at 420s (an orchestration limit, not a model failure). Only ONE draft cleared the 3-pass qwen3.7-flash multipass mean>=8.0 gate: deepseek-v4-flash 8.47 (a=9.00/9.00/9.00 across passes; quality the weak sub-score at 6-7). All seven other survivors failed the gate (gpt-6-luna 7.92, hy3 7.92, minimax-m3 7.85, muse 7.43, mimo 7.38, longcat-2.0 7.27, glm-5.3-flash 7.27). QA: orchestrator mechanical pass - 0 numbered references, 'unlocks' reworded to 'includes' (banned-slop list), HERO_IMAGE_PROMPT stripped, honest scope line present, 10/10 required H2s in order with every self-contained lead measuring 49-59 words (deterministic count), comparison table with the exact contracted columns, FAQ with 4 question-shaped H3s, E-E-A-T 'How this guide was built' block, meta description 152 chars with no digits/percentages, 0 raw angle brackets.

deepseek-v4-flash 8.5

Judge qwen/qwen3.7-flashPlan deepseek-v4-proResearch mimo-v2.6-proQA orchestrator mechanical passIteration 1Threshold met

🥇

deepseek-v4-flashcommandcode

8.5/10

AutoFix/pricing breakdown; flagged all three vendor inconsistencies; comparison table complete; 11 inline citations

Runs long at ~3.0k words, above the runner's ~2500-word cost-control note; quality sub-score 6-7 across passes

🥈

gpt-6-lunacommandcode

7.9/10

Tight 1,957 words; clean prose; 10 external links

Below the 8.0 mean gate (7.92)

🥉

hy3opencode-go

7.9/10

Good structure, table, FAQ

Below gate (7.92); quality sub-score weak on 2 of 3 passes

minimax-m3opencode-go

7.8/10

Solid feature coverage

Below gate (7.85)

muse-spark-1.3-contributoropencode-go

7.4/10

Complete sections

Below gate (7.43)

mimo-v2.6-flashopencode-go

7.4/10

Clean, 0 slop words

Below gate (7.38)

longcat-2.0opencode-go

7.3/10

Complete sections

Below gate (7.27)

glm-5.3-flashopencode-go

7.3/10

Complete sections

Below gate (7.27)