CL-SN55 941 —
GPT-61S 885 -0.6%
GPT-56T 861 —
GPT-6S 825 -1.4%
CL-OP55X 819 -1.4%
MUSE-SPK 793 -1.4%
QWEN-38X 779 -2%
GPT-6A 778 -0.5%
CL-OP5H 712 -1.7%
CL-OP5X 710 -1.7%
KIMI-K3X 700 -2.5%
CL-OP46H 695 -1.7%
CL-FAB5H 693 -1.7%
CL-OP55H 693 -1.7%
CL-OP47H 684 -1.7%
GEM-37FH 649 -1.8%
GEM-38FH 634 -3.4%
GPT-56S 574 -0.5%
CL-OP47 540 -0.6%
INKL 531 —
GEM-3P 494 —
CL-OP46 490 —
CL-OP48 483 —
GEM-31P 459 -5.4%
GPT-6L 436 —
CL-SN55 941 —
GPT-61S 885 -0.6%
GPT-56T 861 —
GPT-6S 825 -1.4%
CL-OP55X 819 -1.4%
MUSE-SPK 793 -1.4%
QWEN-38X 779 -2%
GPT-6A 778 -0.5%
CL-OP5H 712 -1.7%
CL-OP5X 710 -1.7%
KIMI-K3X 700 -2.5%
CL-OP46H 695 -1.7%
CL-FAB5H 693 -1.7%
CL-OP55H 693 -1.7%
CL-OP47H 684 -1.7%
GEM-37FH 649 -1.8%
GEM-38FH 634 -3.4%
GPT-56S 574 -0.5%
CL-OP47 540 -0.6%
INKL 531 —
GEM-3P 494 —
CL-OP46 490 —
CL-OP48 483 —
GEM-31P 459 -5.4%
GPT-6L 436 —
← Back to feed

Google Has Fallen to 10th on the Artificial Analysis Intelligence Index With No Flagship Model in Sight

Google’s position on the Artificial Analysis Intelligence Index has slipped to 10th place, with 13 models from at least eight rival labs now ranked above its best available model. The gap is no longer just Anthropic and OpenAI: Chinese labs including Z.AI, Alibaba, Moonshot, Step, and DeepSeek have each placed models above Google in the index.

As of September 22, 2026 — one week before Claude Opus 5.5 and Claude Fable 5.1 registered their latest scores — the index showed Claude Fable 5.1 (max, with tools) at 53, Claude Opus at 51, Meta’s Muse Spark 1.3 at 48, and Grok 4.7 at 46. Alibaba’s Qwen3.8 Max and Z.AI’s GLM-5.3 both scored 45. A three-way tie at 44 included Grok 4.6, Step 3 Preview, and Moonshot’s Kimi K3.

Google sat at 10th, 13 points below the leaders, with 13 models from at least eight labs between it and first place.

The Flash-Only Strategy

The immediate cause is product: Google’s entire 2026 model strategy has been executed through the Flash line. Gemini 2.5 Flash, Gemini 3 Flash, and a series of iterative point releases have shipped at a fast cadence while a Gemini Pro-tier successor to Gemini 3.1 Pro has not appeared on the index. Gemini 3.1 Pro Preview, the lab’s most capable model on the standard benchmark circuit, scored 80.6% on SWE-bench Verified — below DeepSeek V4 Pro (96.4%), Claude Fable 5 (95.0%), and Kimi K3 (93.4%).

Gemini 4.0, which internal and analyst reports have referenced as Google’s next flagship, has not been released or scored on external benchmarks. Reports cited by Artificial Analysis suggest it remains some time off. The gap between Gemini 3.1 Pro Preview’s last external score and the frontier has widened with every model release from competing labs.

Who Is Above Google

The roster of labs with models above Google on the Intelligence Index as of September 22 is notable for its breadth:

  • Anthropic — Claude Fable 5.1 (53), Claude Opus 5.5 (51)
  • OpenAI — Multiple frontier variants
  • xAI — Grok 4.7 (46), Grok 4.6 (44)
  • Meta — Muse Spark 1.3 (48)
  • Z.AI — GLM-5.3 (45)
  • Alibaba — Qwen3.8 Max (45)
  • Moonshot AI — Kimi K3 (44)
  • StepFun — Step 3 Preview (44)
  • DeepSeek — Cited in the same cohort

The list includes five Chinese AI labs, a shift from early 2026 when the frontier was effectively a three-lab race between Anthropic, OpenAI, and Google. Since then, Chinese model quality has compressed the gap significantly. Stack Futures tracked this milestone in June when China held all 10 spots on the open-weights sub-index; the closed API rankings have followed on a lag.

What Comes Next

Google’s next genuine opportunity to climb the index is Gemini 4.0. No release date has been announced and no benchmark figures have been published. Historically, Gemini releases have been accompanied by GPQA Diamond, SWE-bench, and MMLU scores. Until those appear, the gap will continue widening as competing labs ship new models.

The Artificial Analysis Intelligence Index is updated continuously from external benchmark results rather than self-reported scores. Models listed reflect scores from independent evaluators, not lab-published claims. Current first place as of the September 27 scrape — following Claude Opus 5.5’s September 22 launch — is Claude Opus 5.5 max, which overtook Fable 5.1 on the Artificial Analysis ranking after Opus 5.5 became available.