Grok 4.5

+ contribute

Grok 4.5 is a closed-weights language model from SpaceXAI (xAI), released July 2026.

75.0±0.7#41 of 719 LLMsfrom 20 benchmark results across 7 sources

$ how it compares

The LLMs ranked either side of Grok 4.5, out of 828 with a score, then more from SpaceXAI (xAI).

#	model	score	where	released
40	Muse Spark 1.2 Meta	75±1	api $2.00/M	2026-08
41	Claude Opus 4.7 Anthropic	75±1	api $10.00/M	2026-04
42	Grok 4.5 SpaceXAI (xAI)	75±1	api $3.00/M	2026-07
43	Muse Spark 1.1 Meta	75±1	api $2.00/M	2026-07
44	Muse Spark Meta	75±2	comparison only	2026-04
	more from SpaceXAI (xAI)
30	Grok 4.6 SpaceXAI (xAI)	77±1	api $3.00/M	2026-08
47	Grok 4.7 SpaceXAI (xAI)	75±1	api $3.00/M	2026-09
60	Grok 4.20 Multi-Agent (0309) SpaceXAI (xAI)	73±2	api $1.56/M	2026-03

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #38 of 300 · 16% of its score · 4 tests

Beats Gemini 3.7 Flash, Qwen3.8-Flash-Next, Qwen3.8-Omni-Flash here, though they rank higher overall. Trails GLM-5.3-Flash, Grok 4.7, Claude Sonnet 5 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena Agent (October 2026 aggregate IPS)	0.0121 IPS ✎	▁▁▁▄▂▃▇█▂▂▁▃ beats 24/49	avg -0.312 IPS	▇▇▇ 8.0%	sharp
Terminal-Bench 4.0	12.4% ✎	█▅▅▅▅·▅▃▃▅▅▅ beats 2/20	very hard · avg 0.4%	▇▇ 4.0%	sharp
τ³-bench banking_knowledge (v1.0.1 grading)	47.9% ✎	▃▃▂▃·▆▃▂█▅▅▂ beats 23/25	very hard · avg 4.1%	▇ 3.0%	at its level
EEGAgentBench overall (tool-based EEG analysis agents, 6 tasks)	51.8% ✎	▂·▁▁··▇▇█▂·▁ beats 22/25	hard · avg 29%	▇ 1.2%	blunt · at its level

coding▁▁▁▃▆██▇▆▄▃▁ #36 of 458 * · 10% of its score · 2 tests

Beats GPT-5.5, GPT-5.6 Terra, GPT-5.4 here, though they rank higher overall. Trails Qwen3.8 27B, GLM-5.2, GLM-5.3-Flash here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena WebDev (Code Arena)	1551.8 ✎	▁▂▄▃▅█▄▇▆▄▁▁ beats 85/119	avg 981	▇▇▇ 7.1%	sharp
Luau Leetcode fully solved (Torpedo Software)	78.4% ✎	▁▁▁▁·▁▂▃▅▃█▇ beats 30/49	hard · avg 18%	▇ 2.8%	at its level

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #37 of 393 * · 4.0% of its score · 1 test

Beats GPT-6 Sol, MiMo-V2.6-Flash (309B-A15B), GPT-6 Luna here, though they rank higher overall. Trails MiMo-V2.5-Pro (1T-A42B), Grok 4.20 (0309 reasoning), Claude Opus 4.5 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
LMArena (text)	1465.8 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 338/379	avg 1268	▇▇ 4.0%	blunt

math▁▂▃▅▆▅▄█▄▁▁▁ #36 of 398 · 48% of its score · 5 tests

Beats Claude Opus 4.7 here, though they rank higher overall. Trails GPT-5.2, Qwen3.7 Max, Claude Sonnet 5 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
FrontierMath tiers 1–3 v2	57.2% ✎	▂··▁▁▃▄▄█▅▃▃ beats 43/82	very hard · avg 1.2%	▇▇▇▇▇▇▇ 19%	sharp · at its level
FrontierMath tier 4 v2	24.4% ✎	▁▂▁▂█▅▂▂▁▁▁▁ beats 25/63	very hard · avg 0.1%	▇▇▇▇▇▇ 16%	sharp · at its level
MathArena ArXivMath June 2026 (48 problems)	66.7% ✎	▂···▄·██▆▆▄▂ beats 8/19	very hard · avg 4.6%	▇▇▇▇ 11%	sharp · at its level
MathArena BrokenArXiv June 2026 (54 problems)	29.3% ✎	▄··█▂▃▂▆▄··▂ beats 7/14	very hard · avg 1.1%	▇ 2.5%	sharp · at its level · noisy
OTIS mock AIME 2024–2025	97.8% ✎	▃▃▄▃▃▃▇▇█▃▃▅ beats 171/193	very hard · avg 8.5%	▇ 0.8%	sharp · maxed out · noisy

vision & documents▁▁▃▂▁▃▄▂▆▇█▄ #26 of 147 * · 4.1% of its score · 2 tests

Beats Claude Sonnet 5.5, MiMo-V2.6-Pro, GPT-5.6 Terra here, though they rank higher overall. Trails Gemini 3.6 Flash, Gemini 3.5 Flash, GLM-5.3-Flash here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena vision (style controlled)	1279.0 ✎	▁▁▂▂▂▂▄▅▅▆██ beats 123/145	avg 1068	▇ 2.8%	blunt
Arena document (style controlled)	1464.3 ✎	▁▁▃▃▄▃▇█▅▇▄▅ beats 20/37	avg 1335	▇ 1.3%	blunt

web search▂▁▂▆▄█▃▃▁▁·▁ #27 of 80 * · 1.6% of its score · 1 test

Beats Claude Haiku 5.5, Gemini 3.1 Pro (Preview), Claude Opus 4.7 here, though they rank higher overall. Trails GPT-4o mini (2024-07-18), Mistral Small 3.1 24B (2503), GPT-4o (2024-08-06) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena search systems (style controlled)	1202.4 ✎	▁·▁▁▂▂▂█▅▂▁▁ beats 23/30	avg 1055	▇ 1.6%	blunt

other tests5 not in a task · 16% of its score

test	result	vs the others in its run	difficulty	counts	why
Furniture Assembly (Epoch)	22.5% ✎	█▆▆▆▆▂▅▂▂▂·▅ beats 3/30	very hard · avg 1.2%	▇▇▇ 6.6%	sharp · at its level
Chess Puzzles (Epoch)	36.0% ✎	█▁▁▁▁▂▁▄▃▂▁▁ beats 112/135	very hard · avg 5.9%	▇ 3.3%	sharp · at its level · noisy
GPQA Diamond	93.4% ✎	▁▁▁▁▂▄▆▄▅▇█▅ beats 197/208	hard · avg 40%	▇ 3.1%	sharp
SimpleQA Verified	48.3% ✎	▁▁▃▂▃▃▆▆█▃▂▅ beats 50/78	very hard · avg 8.0%	▇ 1.6%	at its level
RobloxQA v2.0 multiple choice (Torpedo Software)	96.2% ✎	▁▁▃▁█▄▃▃▆▂▂▂ beats 33/49	easy · avg 78%	▇ 0.9%	maxed out

$ where to run it

Through an API

api · xAI · $3.00/M · $2 in / $6 out per 1M tokens · 500k context · docs.x.ai ↗ · 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… 50%, LMArena leaderboard dataset 25%, MathArena ArXivMath / BrokenA… 13% and 4 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	7	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 50%	CC BY 4.0
LMArena leaderboard dataset	6	▇▇▇▇▇▇▇ 25%	CC BY 4.0
MathArena ArXivMath / BrokenArXiv / ArXivLean model-output datasets	2	▇▇▇▇ 13%	CC BY-SA 4.0
Terminal-Bench 4.0 official leaderboard submissions	1	▇ 4%	Apache-2.0
Roblox LLM Leaderboard (Torpedo Software): RobloxQA v2.0 and Luau Leetcode	2	▇ 4%	MIT
τ²/τ³-bench leaderboard: Sierra's own evaluation runs	1	▇ 3%	MIT
EEGAgentBench: Benchmarking LLM Agents on Short- and Long-Horizon EEG Analysis (arXiv:2609.31632v1)	1	▇ 1%	CC BY 4.0
prior · pull toward the middle		▇ 0%
weighted average		= 75.0