SI Index Data as of Oct 8, 2026

Gemini 3 Pro Preview: Price, Context, Benchmarks, and Release Details

Google provisional listing · first seen Oct 8, 2026 · score computed Oct 8, 2026, 22:13 UTC
56.8
SI Score (method si-v2-absolute-shrinkage-1)
#67 of 115 ranked
53% confidence 53 percent, Medium confidence — 3 of 7 expected sources in
Coverage 43% of expected source weight · 100% confidence at 80% coverage

Pillar breakdown

Coding (weight 40 percent) 63.7
Math (weight 15 percent) 91.4
Preference (weight 15 percent) —
Reasoning (weight 30 percent) 55.9

Pillar weights: reasoning 30% · math 15% · coding 40% · preference 15%. Benchmark results use fixed 0–100 scales before averaging; incomplete evidence is shrunk toward 50.

Facts

Input price / 1M
not yet reported
Output price / 1M
not yet reported
Context window
1Mmodels.devPublished source fact Retrieved Oct 8, 2026 · MIT
Open source ↗
Max output
65.5Kmodels.devPublished source fact Retrieved Oct 8, 2026 · MIT
Open source ↗
Released
Nov 18, 2025models.devPublished source fact Retrieved Oct 8, 2026 · MIT
Open source ↗
Open weights
Nomodels.devPublished source fact Retrieved Oct 8, 2026 · MIT
Open source ↗
License
not yet reported
Input modalities
text, image, video, audio, pdfmodels.devPublished source fact Retrieved Oct 8, 2026 · MIT
Open source ↗

Benchmark results

8 results
Benchmark Raw result Normalized (0–100) Pillar
gpqa diamond 92.6%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] Published Nov 19, 2025 Retrieved Oct 8, 2026 · CC-BY
Open source ↗ 		92.6 reasoning
hle scale 37.5%Humanity’s Last ExamPotential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. [variant] Published Nov 19, 2025 Retrieved Oct 8, 2026 · factual citation
Open source ↗ 		37.5 reasoning
otis mock aime 2024 2025 91.4%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] Published Nov 19, 2025 Retrieved Oct 8, 2026 · CC-BY
Open source ↗ 		91.4 math
swe bench pro public 43.3%Official model cards via models.devLab-reported; metric resolve rate; transcribed by MIT models.dev catalog; not independently evaluated [variant] public Retrieved Oct 8, 2026 · factual citation; MIT transcription
Open source ↗ 		43.3 coding
swe bench pro public 43.3%SWE-bench Pro (public)Published steward score [variant] Published Nov 26, 2025 Retrieved Oct 8, 2026 · factual citation
Open source ↗ 		43.3 coding
swe bench verified 72.9%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] Published Feb 13, 2026 Retrieved Oct 8, 2026 · CC-BY
Open source ↗ 		72.9 coding
swe bench verified 77.4%SWE-bench VerifiedSWE-bench published model plus agent result; harness retained, not a base model evaluation [variant] live-SWE-agentPublished Nov 20, 2025 Retrieved Oct 8, 2026 · factual citation
Open source ↗ 		77.4 coding
swe bench verified 74.2%SWE-bench VerifiedSWE-bench published model plus agent result; harness retained, not a base model evaluation [variant] mini-SWE-agent; 1.15.0Published Nov 18, 2025 Retrieved Oct 8, 2026 · factual citation
Open source ↗ 		74.2 coding

Normalization uses fixed absolute 0–100 scales for each unit, independently of other models that have a result on each benchmark. Coverage and evidence breadth still affect the composite; compare the evaluation conditions before reading a small score gap as decisive. “lab-reported” marks the provider's own published figure.

Sources: in and pending

in Reported (3)

  • Epoch AI Benchmarking · arrived Oct 8, 2026
  • Humanity’s Last Exam · arrived Oct 8, 2026
  • Official model cards via models.dev · arrived Oct 8, 2026

pending Awaiting (4)

  • ARC Prize · carries 13% of expected weight
  • LiveBench · carries 13% of expected weight
  • LMArena / Arena · carries 25% of expected weight
  • Terminal-Bench · carries 6% of expected weight

The confidence % rises as pending sources publish. Some sources never cover some models — that is why 100% confidence arrives at 80% of expected weight, not at full coverage.