Mindshare · ModelsLive · 30 models · GPT-5 Chat leads at 18.5%
Model Mindshare · Tracked
Which AI models dominate the conversation? Every model ranked by attention across Reddit, Twitter, HackerNews, GitHub, arXiv, and news. Updated daily.
GPT-5 Chat · 18.5%Claude Mythos Preview · +520bps6 sources trackedRefreshed daily
Model Mindshare Arena
PROPORTIONAL BLOCKS BY MINDSHARE · BIGGER = MORE ATTENTION
Top Gainer
·
Top Gainers
Claude Mythos Preview14.2%+180bps+520bps
DeepSeek V3.2 Speciale6.2%+90bps+280bps
Qwen3.5 397B5.8%+60bps+210bps
MiMo-V2-Flash1.8%+40bps+180bps
Step 3.5 Flash2.1%+35bps+120bps
Phi-50.9%+25bps+90bps
Mistral Large 31.0%+10bps+30bpsTop Losers
GPT-5 Chat18.5%-120bps-340bpsClaude Opus 4.68.5%-30bps-180bpsGPT-4o4.5%-40bps-150bps
Grok 33.8%-15bps-90bps
Gemini 3.1 Pro11.8%+45bps-80bps
Llama 44.2%+20bps-60bpso4 Mini1.5%-20bps-60bpsGPT-5 ChatPricing backlash on Twitter
18.5%
↓ -340bpsClaude Mythos PreviewSafety benchmarks impressed researchers on arXiv
14.2%
↑ +520bpsGemini 3.1 ProSteady enterprise adoption, low controversy
11.8%
↓ -80bpsClaude Opus 4.6Attention cannibalized by Mythos Preview launch
8.5%
↓ -180bpsDeepSeek V3.2 Speciale6.2%
↑ +280bpsQwen3.5 397B5.8%
↑ +210bpsGPT-4o4.5%
↓ -150bpsLlama 44.2%
↓ -60bpsGrok 33.8%
↓ -90bpsCInstant
2.5%
Step 3.5 Flash
2.1%
MiMo-V2-Flash
1.8%
o4 Mini
1.5%
Gem 2.5 Flash
1.3%
Mistral Large 3
1.0%
Phi-5
0.9%
Command R+ 20.8%
CSonnet 4.60.8%
Gemma 30.7%
Yi-2 200B0.6%
BenchGecko.ai
BenchGecko.ai·Model Mindshare Arena
Model GMI Leaderboard
30 models · ranked by gecko mindshare index
#ModelGMIShare1D7DSentimentSources
189/10018.5%↓ -120bps↓ -340bps
GPT-5 Chat
GMI 89/100 · 18.5%
-340bps
292/10014.2%↑ +180bps↑ +520bps
Claude Mythos Preview
GMI 92/100 · 14.2%
+520bps
385/10011.8%↑ +45bps↓ -80bps
Gemini 3.1 Pro
GMI 85/100 · 11.8%
-80bps
482/1008.5%↓ -30bps↓ -180bps
Claude Opus 4.6
GMI 82/100 · 8.5%
-180bps
578/1006.2%↑ +90bps↑ +280bps
DeepSeek V3.2 Speciale
GMI 78/100 · 6.2%
+280bps
674/1005.8%↑ +60bps↑ +210bps
Qwen3.5 397B
GMI 74/100 · 5.8%
+210bps
765/1004.5%↓ -40bps↓ -150bps
GPT-4o
GMI 65/100 · 4.5%
-150bps
872/1004.2%↑ +20bps↓ -60bps
Llama 4
GMI 72/100 · 4.2%
-60bps
970/1003.8%↓ -15bps↓ -90bps
Grok 3
GMI 70/100 · 3.8%
-90bps
1055/1002.5%↓ -10bps↓ -45bps
Claude Instant
GMI 55/100 · 2.5%
-45bps
1152/1002.1%↑ +35bps↑ +120bps
Step 3.5 Flash
GMI 52/100 · 2.1%
+120bps
1248/1001.8%↑ +40bps↑ +180bps
MiMo-V2-Flash
GMI 48/100 · 1.8%
+180bps
1358/1001.5%↓ -20bps↓ -60bps
o4 Mini
GMI 58/100 · 1.5%
-60bps
1450/1001.3%↓ -10bps↓ -50bps
Gemini 2.5 Flash
GMI 50/100 · 1.3%
-50bps
1556/1001.0%↑ +10bps↑ +30bps
Mistral Large 3
GMI 56/100 · 1.0%
+30bps
1645/1000.9%↑ +25bps↑ +90bps
Phi-5
GMI 45/100 · 0.9%
+90bps
17
42/1000.8%↑ +5bps↑ +20bps
Command R+ 2
GMI 42/100 · 0.8%
+20bps
1860/1000.8%↓ -5bps↓ -30bps
Claude Sonnet 4.6
GMI 60/100 · 0.8%
-30bps
1944/1000.7%↓ -5bps↓ -20bps
Gemma 3
GMI 44/100 · 0.7%
-20bps
20
38/1000.6%→ +0bps↓ -15bps
Yi-2 200B
GMI 38/100 · 0.6%
-15bps
21
35/1000.5%↑ +5bps↑ +15bps
Jamba 2
GMI 35/100 · 0.5%
+15bps
22
36/1000.5%→ +0bps↑ +10bps
DBRX 2
GMI 36/100 · 0.5%
+10bps
23
32/1000.4%↑ +10bps↑ +40bps
Reka Flash 3
GMI 32/100 · 0.4%
+40bps
2430/1000.4%↑ +5bps↑ +20bps
I3
InternLM 3
GMI 30/100 · 0.4%
+20bps
2528/1000.3%↑ +5bps↑ +15bps
Aya 3
GMI 28/100 · 0.3%
+15bps
26
26/1000.3%→ +0bps↓ -10bps
Falcon 3
GMI 26/100 · 0.3%
-10bps
2725/1000.3%↑ +5bps↑ +10bps
G5
GLM-5
GMI 25/100 · 0.3%
+10bps
28
24/1000.2%→ +0bps↑ +5bps
MiniMax-Text-02
GMI 24/100 · 0.2%
+5bps
29
30/1000.2%↑ +5bps↑ +20bps
Nemotron 5
GMI 30/100 · 0.2%
+20bps
3040/1000.2%↓ -5bps↓ -10bps
Claude Haiku 4
GMI 40/100 · 0.2%
-10bps
Hype vs Reality
mindshare vs benchmark performance · the scatter that matters
Overhyped
Deserved Leaders
Quiet
Underrated
Benchmark Score →
↑ Mindshare
GPT-5 ChatGPT-5 Chat
Claude Mythos PreviewClaude Mythos Preview
Gemini 3.1 ProGemini 3.1 Pro
Claude Opus 4.6Claude Opus 4.6
DeepSeek V3.2 SpecialeDeepSeek V3.2 Speciale
Qwen3.5 397BQwen3.5 397B
GPT-4oGPT-4o
Llama 4Llama 4
Grok 3Grok 3
Claude InstantClaude Instant
Step 3.5 Flash
MiMo-V2-Flash
o4 Mini
Gemini 2.5 Flash
Mistral Large 3
Phi-5
Command R+ 2
Claude Sonnet 4.6
Gemma 3
Yi-2 200B
Jamba 2
DBRX 2
Reka Flash 3
InternLM 3
Aya 3
Falcon 3
GLM-5
MiniMax-Text-02
Nemotron 5
Claude Haiku 4
OverhypedDeservedUnderratedQuiet
Sentiment Breakdown
positive · negative · neutral per model
GPT-5 Chat
35% pos · 42% neg
Claude Mythos Preview
71% pos · 8% neg
Gemini 3.1 Pro
52% pos · 18% neg
Claude Opus 4.6
65% pos · 10% neg
DeepSeek V3.2 Speciale
68% pos · 5% neg
Qwen3.5 397B
62% pos · 8% neg
GPT-4o
45% pos · 22% neg
Llama 4
55% pos · 15% neg
Grok 3
40% pos · 30% neg
Claude Instant
58% pos · 12% neg
Step 3.5 Flash
60% pos · 8% neg
MiMo-V2-Flash
55% pos · 10% neg
o4 Mini
48% pos · 18% neg
Gemini 2.5 Flash
50% pos · 15% neg
Mistral Large 3
55% pos · 12% neg
Positive Neutral Negative
Model Mindshare FAQ
The Gecko Mindshare Index tracks what percentage of total AI attention each model, agent, company, and person commands across Reddit, Twitter/X, Hacker News, GitHub, arXiv, and news outlets. A GMI of 92 means the entity is in the elite tier of attention. All weights are published at /mindshare/methodology.