Gemini 4 Carbon
Top intelligence score here at 60, but restricted to vetted cyber-defence organisations
Benchmark scores
How it is measured →Ranked 1 · bars are scaled to the field maximum for that benchmark.
| Benchmarks | Score | Vs. best in field | Gap |
|---|---|---|---|
| AA Intelligence Index v5 Intelligence index | 60.0 | 100% | |
| AA Intelligence Index v4.3.2 | — | — | not indexed |
| GPQA Diamond | — | — | not indexed |
| Terminal-Bench 4.0 Terminal-agent tasks | 65.2 | 97% | −2.0 pts |
| LiveBench Design | — | — | not indexed |
| Vals Index | — | — | not indexed |
Assessment
Highest AA Intelligence Index v5 score (60, re-scored under v5.0.1), second on Terminal-Bench 4.0 (65.2%) behind Kimi K3.1 Pro, and first on AA-Omniscience at 53.5. Access is restricted, though: only vetted organisations in Google's Fairwind programme can use it, with no public API or App timeline. Raw factual accuracy is 58% and its 11% hallucination rate is higher than Argon's 4%. The $2/$10 launch promo has ended; the regular price is now $4/$20 ($0.2 cache hit).
At the same price
The four models closest on blended price, side by side.
| Model | v5 | Blended | tok/s | Score gap |
|---|---|---|---|---|
| Claude Opus 5.5Anthropic | 59.1 | $8.00 | 94 | -0.9 |
| Muse WonderMeta | 44.0 | $8.00 | 143 | -16.0 |
| Claude Sonnet 5.5Anthropic | 54.5 | $4.00 | 135 | -5.5 |
| GPT-6.1 SolOpenAI | 54.6 | $4.00 | 77 | -5.4 |
Compare side by side
Line it up against other models on price, speed and every benchmark.
The comparison tool lines up to 4 models side by side with switchable normalisation — built for purchasing decisions.
Open the comparison tool → Back to leaderboard