Skip to content

Gemini 4 Neon vs Claude Opus 5.5: Neon Beats Opus 5.5?

Google has officially announced Gemini 4 Neon, and on Google's own benchmarks it beats Claude Opus 5.5 on most tests. On the Artificial Analysis Intelligence Index, Neon scores 61.4 at xhigh against 59.1 for Opus 5.5 at max, and it costs about 36% less per task. Opus still wins a few, including Terminal-Bench 4.1, FrontierCode and AA-Briefcase. Here are the detailed benchmarks to help you see the real differences between these models.

Google's benchmarks

Benchmark
Gemini 4 Neon
Claude Opus 5.5
Difference
AA Intelligence Index v5.0.16159Gemini 4 Neon+2
Vals Index v2.270.9%70.1%Gemini 4 Neon+0.8 pts
CursorBench 4.165.4%58.1%Gemini 4 Neon+7.3 pts
Terminal-Bench 4.165.2%65.8%Claude Opus 5.5+0.6 pts
FrontierCode 2.046.2%47.3%Claude Opus 5.5+1.1 pts
SWE-sweep (provider harness)39.0%10.1%Gemini 4 Neon+28.9 pts
SWE-sweep (Mini-SWE harness)12.8%3.6%Gemini 4 Neon+9.2 pts
AutomationBench56.3%42.5%Gemini 4 Neon+13.8 pts
Agents' Last Exam v1.175.8%71.6%Gemini 4 Neon+4.2 pts
RiemannBench84.0%80.0%Gemini 4 Neon+4.0 pts
Terminal-Bench Science 0.170.5%63.3%Gemini 4 Neon+7.2 pts
HLE-Diamond71.3%67.9%Gemini 4 Neon+3.4 pts
MKA (Climb WSR)56.5%53.0%Gemini 4 Neon+3.5 pts
Chartography73.5%66.3%Gemini 4 Neon+7.2 pts
LVBench89.0%83.7%Gemini 4 Neon+5.3 pts
GraphWalks (256K to 1M)95.2%66.8%Gemini 4 Neon+28.4 pts
CWE-Bench v270.7%67.1%Gemini 4 Neon+3.6 pts
WebExploitBench v1.251.3%37.1%Gemini 4 Neon+14.2 pts
SEC-bench Pro90.5%Not reported

Artificial Analysis benchmarks at the highest effort

Intelligence Index v5.0.161.459.1Gemini 4 Neon+2.3
AA-Briefcase v1.11700 Elo1800 EloClaude Opus 5.5+100 Elo
AutomationBench-AA79.3%69.5%Gemini 4 Neon+9.8 pts
Terminal-Bench 4.063.6%59.6%Gemini 4 Neon+4.0 pts
FrontierCode-AA 2.044.4%46.4%Claude Opus 5.5+2.0 pts
AA-Omniscience v1.154.848.1Gemini 4 Neon+6.7
GDP.pdf28.0%26.2%Gemini 4 Neon+1.8 pts
HLE-Diamond71.3%67.9%Gemini 4 Neon+3.4 pts
CritPt v2 (Hard Subset)60.5%68.0%Claude Opus 5.5+7.5 pts
Terminal-Bench Science 0.170.0%59.0%Gemini 4 Neon+11.0 pts

Intelligence Index by reasoning effort

MaxNo max level59
Xhigh6157Gemini 4 Neon+4
High6052Gemini 4 Neon+8
Medium5548Gemini 4 Neon+7
Low4435Gemini 4 Neon+9

Cost per Intelligence Index task by reasoning effort

MaxNo max level$10.04
Xhigh$6.44$6.05Claude Opus 5.56% cheaper
High$3.74$3.46Claude Opus 5.57% cheaper
Medium$1.90$2.45Gemini 4 Neon22% cheaper
Low$0.75$0.98Gemini 4 Neon23% cheaper

Speed, cost and tokens at the highest effort

Cost per Intelligence Index task$6.44$10.04Gemini 4 Neon36% cheaper
Time per task21.5 min24.5 minGemini 4 Neon12% faster
Output tokens across the Index283M218MClaude Opus 5.523% fewer
Output speed110 tokens/s118 tokens/sClaude Opus 5.57% faster
Time to first answer token56.5 s336 sGemini 4 Neon83% lower

Pricing and specs

Input, per 1M tokens$3.00$4.00Gemini 4 Neon25% cheaper
Output, per 1M tokens$15.00$20.00Gemini 4 Neon25% cheaper
Cached input, per 1M tokens$0.15$0.20Gemini 4 Neon25% cheaper
Batch input, per 1M tokens$1.50$2.00Gemini 4 Neon25% cheaper
Batch output, per 1M tokens$7.50$10.00Gemini 4 Neon25% cheaper
Context window1.05M tokens, input and output combined1M tokens
Standard max output64K tokens128K tokensClaude Opus 5.52x larger
Extended outputUp to 1M with long decode300K in batch, beta
Reasoning levelsLow, medium, high and xhighLow to max with adaptive thinking
Text and image inputYesYes
Audio inputYesNo
Video inputYesNo
PDF inputYesYes
Computer useYesYes
Code executionYesYes
Free API tierNoNo
Knowledge cutoffMarch 2026June 2026Claude Opus 5.53 months newer
ReleasedNovember 18, 2026September 22, 2026Gemini 4 Neon57 days newer