[whatmodel.sh](../../README.md) / Gemini 3.8 Flash
# Gemini 3.8 Flash
Google · released 2026-09 · closed weights
**score 75.5** ±0.5 · #28 of 732 · from 33 benchmark results across 11 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Gemini 3.8 Flash among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃▆█▄▅▅▄●▂▁ #25 of 313 21% 9
coding ▁▁▁▃▅▇█▇▅▄●▁ #35 of 469 12% 5
chat & writing ▁▁▁▃▆▄▆█▆▇▇● #9 of 402* 3.0% 1
math ▁▁▃▅█▆▄█●▂▁▁ #35 of 431 45% 10
vision & documents ▁▁▂▂▁▃▄▂▆▇█● #6 of 153* 4.7% 2
agent tasks · #25 of 313 · 21% of its score
Beats [Hy4 preview (770B-A49B)](hy4-preview.md), [GPT-5.6 Terra](gpt-5-6-terra.md), [Qwen3.8 Max (2.4T-A95B)](qwen3-8-2-4t-a95b.md) here, though they rank higher overall. Trails [Grok 4.7](grok-4-7.md), [Claude Haiku 5.5](claude-haiku-5-5.md), [DeepSeek V4.1 Flash](deepseek-v4-1-flash.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arena Agent (October 2026 aggrega… 0.0265 IPS ▁▁▁▃▃▂●▁▂▂▁▁ beats 39/59 avg -0.379 IPS 7.6% sharp
Terminal-Bench 4.0 19.1% █▅●▅▅·▅▃▃▅▅▅ beats 6/20 very hard · avg 0.5% 3.1% sharp · at its level
Terminal-Bench 2.1 (maker) 90.8% ▁·▁·▁▁▄▃▆█▅● beats 63/63 hard · avg 25% 3.0% sharp · small run · several runs
SWE-bench Pro (split unspecified)… 61.6% ▂·▁▁▁▁▂▃█●▂▁ beats 45/58 very hard · avg 6.0% 1.8% sharp · at its level · small run
E2E-SWE (build a working codebase… 36.7% ▃··▅··●█▃▅▃▅ beats 3/12 very hard · avg 3.7% 1.8% sharp · at its level
Video2World V2WScore (coding agen… 30.1% █··●█·▄····▄ beats 2/6 very hard · avg 7.3% 1.2% at its level
KlinikeBench strict pass@1 (333 i… 18.9% ▁▁·▁▁▁▁▅▃█●█ beats 18/27 very hard · avg 2.4% 0.9% typical
DeepSWE v1.1 (resolved) (maker) 73.7% ▂····▂▁▁▆██● beats 25/31 hard · avg 25% 0.9% at its level · small run
τ³-bench Banking (maker) 38.1% ▂▂·▅█▂▃▃▃▆▂● beats 24/24 very hard · avg 7.9% 0.4% blunt · at its level · small run
coding · #35 of 469 · 12% of its score
Beats [GPT-5.5](gpt-5-5.md), [GPT-5.6 Terra](gpt-5-6-terra.md), [Claude Opus 4.8](claude-opus-4-8.md) here, though they rank higher overall. Trails [GLM-5.2](glm-5-2.md), [GLM-5.3-Flash](glm-5-3-flash.md), [Grok 4.7](grok-4-7.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arena WebDev (Code Arena) 1583.3 ▁▁▃▂▅█▃▆●▅▁▂ beats 96/129 avg 981 5.3% sharp
Luau Leetcode fully solved (Torpe… 84.6% ▁▁▁▁·▁▂▃▅▃█● beats 44/49 hard · avg 17% 1.9% sharp
SWE-bench Pro (split unspecified)… 61.6% ▂·▁▁▁▁▂▃█●▂▁ beats 45/58 very hard · avg 6.0% 1.8% sharp · at its level · small run
E2E-SWE (build a working codebase… 36.7% ▃··▅··●█▃▅▃▅ beats 3/12 very hard · avg 3.7% 1.8% sharp · at its level
DeepSWE v1.1 (resolved) (maker) 73.7% ▂····▂▁▁▆██● beats 25/31 hard · avg 25% 0.9% at its level · small run
chat & writing · #9 of 402 · 3.0% of its score
Beats [GPT-6 Astra](gpt-6-astra.md), [GPT-6.1 Sol](gpt-6-1-sol.md), [Claude Sonnet 5.5](claude-sonnet-5-5.md) here, though they rank higher overall. Trails [Claude Opus 4.6](claude-opus-4-6.md), [Claude Opus 4.7](claude-opus-4-7.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LMArena (text) 1495.9 ▁▁▁▂▃▃▃▆▆▇█● beats 377/385 avg 1260 3.0% blunt
math · #35 of 431 · 45% of its score
Trails [GPT-5.2](gpt-5-2.md), [Qwen3.7 Max](qwen3-7-max.md), [Gemini 3.1 Pro (Preview)](gemini-3-1-pro.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
FrontierMath tier 4 v2 22.0% ▁▂▁▂●▅▂▃▂▁▁▁ beats 24/67 very hard · avg 0.1% 13% sharp · at its level
FrontierMath tiers 1–3 v2 68.4% ▂··▁▁▃▄▄█●▄▃ beats 57/86 very hard · avg 1.2% 12% sharp · at its level
MathArena ArXivMath June 2026 (48… 67.4% ▂···▄·█●▆▆▄▂ beats 10/19 very hard · avg 5.1% 6.1% sharp · at its level
MathArena ArXivMath April 2026 (4… 64.2% ▁·····▅█●▄·▁ beats 16/20 very hard · avg 7.6% 5.9% sharp · at its level
MathArena BrokenArXiv April 2026 … 33.6% ▂▂▆█●▃▂▂▂··▂ beats 11/18 very hard · avg 0.8% 2.4% sharp · at its level · noisy
MathArena ArXivMath May 2026 (40 … 58.3% ▁··▁▄●▄▁·▃·▁ beats 10/17 very hard · avg 11% 2.0% at its level
MathArena ArXivLean June 2026 (46… 19.6% ▄····▄▄●██·▄ beats 4/9 very hard · avg 0.9% 1.8% sharp
MathArena ArXivMath August 2026, … 40.4% ●▄·▄·······█ beats 1/5 very hard · avg 5.9% 1.2% sharp · at its level
MathArena BrokenArXiv August 2026… 27.4% ▄●·····▄·▄·█ beats 1/5 very hard · avg 6.9% 0.5% at its level · noisy
OTIS mock AIME 2024–2025 98.9% ▃▃▄▃▃▃▇▇█▃▂● beats 181/197 very hard · avg 7.2% 0.3% sharp · maxed out · noisy
vision & documents · #6 of 153 · 4.7% of its score
Beats [GPT-6 Astra](gpt-6-astra.md), [Claude Opus 5.5](claude-opus-5-5.md), [GPT-6.1 Sol](gpt-6-1-sol.md) here, though they rank higher overall. Trails [Claude Opus 4.7](claude-opus-4-7.md), [Gemini 3.7 Flash](gemini-3-7-flash.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
CharXiv (reasoning questions) (ma… 86.2% ▃·▃·▃▃▃·▃▃█● beats 11/11 medium · avg 38% 2.7% typical
Arena vision (style controlled) 1291.0 ▁▁▂▂▂▂▃▄▅▅█● beats 143/152 avg 1061 2.0% blunt
other tests (8)
test result vs the others in its run difficulty counts why
Mystery Game Puzzles (Epoch) 47.0% █··▂▃▇▄█●▄▂▁ beats 64/77 very hard · avg 9.4% 5.3% sharp · at its level · noisy
Furniture Assembly (Epoch) 31.7% █▆●▆▆▂▆▂▂▂▂▅ beats 9/32 very hard · avg 1.2% 5.2% sharp · at its level
Chess Puzzles (Epoch) 61.0% █▁▁▁▁▂▃▃▃▂●▁ beats 135/137 very hard · avg 5.8% 3.0% sharp · at its level · noisy
GPQA Diamond 95.4% ▁▁▁▁▂▄▆▄▅▇█● beats 209/211 hard · avg 39% 1.7% sharp
SimpleQA Verified 69.7% ▁▁▃▂▃▃▆▆█▄▂● beats 75/81 very hard · avg 8.0% 1.2% at its level
Icelandic LLM Leaderboard average… 63.5% ▁▁▃▂▂▄▁▃▄▄●█ beats 28/38 very hard · avg 14% 1.1% at its level
LVBench (maker) 87.1% ▇·█▅▅▄▄··▃●▁ beats 27/28 medium · avg 59% 0.9% typical
RobloxQA v2.0 multiple choice (To… 96.5% ▁▁▃▁█▄▃▃●▂▂▂ beats 35/49 easy · avg 77% 0.7% maxed out
source results share of the score
Epoch AI 8 42% ▇▇▇▇▇
MathArena ArXivMath / BrokenArXiv / ArX… 7 20% ▇▇
LMArena leaderboard dataset 4 18% ▇▇
Google DeepMind Gemini model page and m… 4 6% ▇
Google Cloud developer's guides to Gemi… 3 4% ▇
Terminal-Bench 4.0 official leaderboard… 1 3% ▇
Roblox LLM Leaderboard 2 3% ▇
E2E-SWE: Benchmarking LLMs on Building … 1 2% ▇
Video2World: Benchmarking Coding Agents… 1 1% ▇
Miðeind Icelandic LLM Leaderboard 1 1% ▇
KlinikeBench: Evaluating Language Model… 1 1% ▇
prior: pull toward the middle 0% ▇
weighted average = 75.5
Licences: Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · MathArena ArXivMath / BrokenArXiv / ArXivLean model-output datasets ([CC BY-SA 4.0](../../licenses/matharena-sharealike-NOTICE.txt)) · LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Google Cloud developer's guides to Gemini 3.8 Flash and 3.7 Flash ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · Terminal-Bench 4.0 official leaderboard submissions ([Apache-2.0](../../licenses/terminal-bench-leaderboard-submissions-Apache-2.0.txt)) · Roblox LLM Leaderboard ([MIT](../../licenses/roblox-llm-leaderboard-MIT.txt)) · E2E-SWE: Benchmarking LLMs on Building Working Codebases from Scratch ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Video2World: Benchmarking Coding Agents for Interactive World Modeling from Embodied Videos ([CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/)) · Miðeind Icelandic LLM Leaderboard ([CC BY 4.0](../../licenses/mideind-icelandic-llm-leaderboard-CC-BY-4.0.txt)) · KlinikeBench: Evaluating Language Models Beyond Diagnostic Accuracy ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/))
## How to run it
### Through an API
api (Google) · $1.50/M blended · $0.75 in / $3.75 out per 1M tokens · [ai.google.dev](https://ai.google.dev/gemini-api/docs/models)
[price source](https://github.com/anomalyco/models.dev/blob/a668908cfe7073faa0363699486fc41a3ff52412/providers/google/models/gemini-3.8-flash.toml): Standard base token rate from models.dev. Provider-specific tiers, long-context surcharges, caching and promotions may differ. · [context source](https://github.com/anomalyco/models.dev/blob/a668908cfe7073faa0363699486fc41a3ff52412/models/google/gemini-3.8-flash.toml)
[interactive version](https://whatmodel.sh/model/gemini-3-8-flash) · [JSON](https://whatmodel.sh/api?format=json&all=1)