whatmodel.sh / Gemini 3.6 Flash

Gemini 3.6 Flash

+ contribute

Gemini 3.6 Flash is a closed-weights language model from Google, released July 2026.

73.6±0.6#53 of 719 LLMsfrom 27 benchmark results across 8 sources

$ how it compares

The LLMs ranked either side of Gemini 3.6 Flash, out of 828 with a score, then more from Google.

#	model	score	where	RAM	released
54	GPT-5.2 OpenAI	74±1	api $4.81/M		2025-12
55	Hunyuan Hy3 (295B-A21B) Tencent	74±1	api $0.23/M +local	~96GB	2026-07
56	Gemini 3.6 Flash Google	74±1	api $1.50/M		2026-07
57	GLM-5.2 Z.ai	74±1	api $2.15/M +local	~121GB	2026-06
58	Dola Seed 2.0 Pro ByteDance	74±2	comparison only
	more from Google
9	Gemini 4 Argon (limited preview) Google	82±1	limited preview		2026-09
23	Gemini 3.8 Flash Google	77±1	api $1.50/M		2026-09
32	Gemini 3.7 Flash Google	76±1	api $1.50/M		2026-08

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #70 of 300 · 14% of its score · 4 tests

Trails MiMo-V2.5 (310B-A15B), Muse Glimmer 30B, MiniMax-M3 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena Agent (October 2026 aggregate IPS)	-0.0768 IPS ✎	▁▁▁▄▂▃▇█▂▂▁▃ beats 5/49	avg -0.312 IPS	▇▇▇ 6.5%	sharp
Terminal-Bench 2.1 lab +1	73.0% ✎	▁·▁·▁▁▄▃▄█▃▇ beats 38/56	hard · avg 23%	▇▇ 4.4%	small run · several runs
E2E-SWE (build a working codebase from scratch, 186 tasks) pass@1	11.7% ✎	▃··▅··▃█▃▅▃▅ beats 0/12	very hard · avg 3.1%	▇ 2.3%	sharp · at its level
Terminal-Bench 3.0 lab	5.4% ✎	█·█··██····█ beats 0/4	very hard · avg 1.7%	▇ 0.7%	small run

coding▁▁▁▃▆██▇▆▄▃▁ #41 of 458 * · 8.1% of its score · 2 tests

Beats GPT-5.5, GPT-5.6 Terra, GPT-5.4 here, though they rank higher overall. Trails Mistral Large 4 (public preview), Qwen3.8 27B, GLM-5.2 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena WebDev (Code Arena)	1536.3 ✎	▁▂▄▃▅█▄▇▆▄▁▁ beats 80/119	avg 981	▇▇ 5.8%	sharp
E2E-SWE (build a working codebase from scratch, 186 tasks) pass@1	11.7% ✎	▃··▅··▃█▃▅▃▅ beats 0/12	very hard · avg 3.1%	▇ 2.3%	sharp · at its level

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #16 of 393 * · 3.3% of its score · 1 test

Beats GPT-6 Astra, GPT-6.1 Sol, Claude Sonnet 5.5 here, though they rank higher overall.

test	result	vs the others in its run	difficulty	counts	why
LMArena (text)	1483.0 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 361/379	avg 1268	▇ 3.3%	blunt

math▁▂▃▅▆▅▄█▄▁▁▁ #44 of 398 · 51% of its score · 8 tests

Beats Grok 4.7, GLM-5.3-Flash here, though they rank higher overall. Trails A.X K2 (688B-A33B), Kimi K2.6, GLM-5.2 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
FrontierMath tiers 1–3 v2	58.9% ✎	▂··▁▁▃▄▄█▅▃▃ beats 45/82	very hard · avg 1.2%	▇▇▇▇▇▇ 16%	sharp · at its level
FrontierMath tier 4 v2	22.0% ✎	▁▂▁▂█▅▂▂▁▁▁▁ beats 23/63	very hard · avg 0.1%	▇▇▇▇ 11%	sharp
MathArena ArXivMath June 2026 (48 problems)	56.9% ✎	▂···▄·██▆▆▄▂ beats 6/19	very hard · avg 4.6%	▇▇▇▇ 8.9%	sharp · at its level
MathArena ArXivMath April 2026 (40 problems)	52.5% ✎	▁·····▅██▄·▁ beats 8/20	very hard · avg 7.2%	▇▇▇ 7.4%	sharp · at its level
MathArena ArXivMath May 2026 (40 problems)	54.2% ✎	▁··▁▄█▄▁·▃·▁ beats 7/17	very hard · avg 10%	▇ 2.7%	at its level
MathArena BrokenArXiv April 2026 (61 problems)	16.0% ✎	▂▂▆█▃▃▂▂▂··▂ beats 3/18	very hard · avg 0.7%	▇ 2.6%	sharp · at its level · noisy
MathArena BrokenArXiv June 2026 (54 problems)	12.0% ✎	▄··█▂▂▂▆▄··▂ beats 2/14	very hard · avg 1.1%	▇ 1.9%	sharp · at its level · noisy
OTIS mock AIME 2024–2025	94.2% ✎	▃▃▄▃▃▃▇▇█▃▂▅ beats 158/193	very hard · avg 8.5%	▇ 0.8%	sharp · maxed out · noisy

vision & documents▁▁▃▂▁▃▄▂▆▇█▄ #20 of 147 · 8.1% of its score · 5 tests

Beats Claude Sonnet 5.5, GPT-5.5, MiMo-V2.6-Pro here, though they rank higher overall.

test	result	vs the others in its run	difficulty	counts	why
CharXiv (reasoning questions) lab	85.2% ✎	▃·▃·▃▃▃·▃▃█▅ beats 10/11	medium · avg 40%	▇ 3.6%	typical
Arena vision (style controlled)	1279.7 ✎	▁▁▂▂▂▂▄▅▅▆██ beats 125/145	avg 1068	▇ 2.2%	blunt
SciDocBench scientific-document workflows (overall, 496 instances)	59.9% ✎	▃▃···█▃▅▃·▃▅ beats 9/11	hard · avg 26%	▇ 1.1%	blunt · at its level
Arena document (style controlled)	1453.4 ✎	▁▁▃▃▄▃▇█▅▇▄▅ beats 14/37	avg 1335	▇ 1.1%	blunt
PureDocBench document parsing (Avg3 over Clean/Digital/Real)	78.3% ✎	▂·▁·▂▂▅█·▁▂▃ beats 21/22	medium · avg 62%	▇ 0.1%	blunt · at its level

other tests7 not in a task · 17% of its score

test	result	vs the others in its run	difficulty	counts	why
Furniture Assembly (Epoch)	23.3% ✎	█▆▆▆▆▂▅▂▂▂·▅ beats 4/30	very hard · avg 1.2%	▇▇ 4.9%	sharp · at its level
Mystery Game Puzzles (Epoch)	30.0% ✎	█··▂▃▇▄█▂▃▁▁ beats 49/74	very hard · avg 9.4%	▇ 3.5%	sharp · noisy
GPQA Diamond	94.1% ✎	▁▁▁▁▂▄▆▄▅▇█▅ beats 200/208	hard · avg 40%	▇ 2.9%	sharp
Chess Puzzles (Epoch)	40.0% ✎	█▁▁▁▁▂▁▄▂▃▁▁ beats 119/135	very hard · avg 5.9%	▇ 2.5%	sharp · noisy
SimpleQA Verified	66.2% ✎	▁▁▃▂▃▃▆▆█▃▃▅ beats 67/78	very hard · avg 8.0%	▇ 1.3%	at its level
MRCR v2 8-needle, 128k lab	91.8% ✎	█▄▄█·▄···▄·▄ beats 7/8	medium · avg 47%	▇ 1.2%	small run
LVBench lab	84.2% ✎	▇·█▅▅▄▄··▃▁▁ beats 25/28	medium · avg 57%	▇ 1.2%	typical

$ where to run it

Through an API

api · $1.50/M · $0.75 in / $3.75 out per 1M tokens · 1M context · ai.google.dev ↗ · price sourceStandard base token rate from models.dev. Provider-specific tiers, long-context surcharges, caching and promotions may differ.

$ how the score was worked out · Epoch AI — internal evaluatio… 43%, MathArena ArXivMath / BrokenA… 24%, LMArena leaderboard dataset 19% and 5 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

source	results	share of the score	licence
Epoch AI — internal evaluations	8	▇▇▇▇▇▇▇▇▇▇▇▇▇ 43%	CC BY 4.0
MathArena ArXivMath / BrokenArXiv / ArXivLean model-output datasets	5	▇▇▇▇▇▇▇ 24%	CC BY-SA 4.0
LMArena leaderboard dataset	5	▇▇▇▇▇▇ 19%	CC BY 4.0
Google DeepMind Gemini model page and model cards (lab-reported scores) · lab-reported	5	▇▇▇ 9%	Factual vendor-reported score (not licensed content)
E2E-SWE: Benchmarking LLMs on Building Working Codebases from Scratch (arXiv:2609.38335v1)	1	▇ 2%	CC BY 4.0
Google Cloud developer's guides to Gemini 3.8 Flash and 3.7 Flash · lab-reported	1	▇ 2%	CC-BY-4.0
SciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)	1	▇ 1%	CC BY 4.0
How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)	1	▇ 0%	CC BY 4.0
prior · pull toward the middle		▇ 0%
weighted average		= 73.6