Unverified50% confidenceFactExact time
In the Anti-Gravity CLI test, Gemini 4 Neon spent 56% to 81% of its tokens on thinking, and with no opening filler it gave the answer directly.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/19/2026
First Seen
Valid until: 2/17/2027
Sources
Related Entities
Related Claims
UnverifiedOn SWE-sweep, Gemini 4 Neon scores 12.8% under the mini-SWE harness and 39.0% in Google Antigravity.74% similarUnverifiedCurrent information about Gemini 4 Neon comes from a single RSS newsletter, and its specific parameters, context window, and release timing have not been disclosed; the score of 61 should be regarded as a preliminary signal rather than a final conclusion.70% similarUnverifiedAntigravity (Gemini 3.1 Pro) was the only client in this experiment whose accuracy declined, dropping to 89%, because it skipped reading the schema of the multi-operation tool and guessed the operation name.70% similarUnverifiedAt present, all judgments about the comparison between Gemini 4 Neon and Opus 5.5 are based on the title of a brief post and lack verifiable benchmark data or product details.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1187994API
curl https://kongchang.com/api/v1/knowledge/claims/1187994MCP
get_claim(id=1187994)