GPT-5 nano (high): benchmark scores, price and ranking

GPT-5 nano (high) from OpenAI ranks #661 of 743 LLMs on the LLMs Tiger aggregate benchmark score (43.8), based on 13 benchmark results captured up to 2026-10-13.

Organisation

OpenAI

Type

Fast / low-cost (flash) model

Open weights

No

Released

2025-08-07 (source: curated)

Cheapest API price

$0.050 in / $0.400 out per million tokens (via azure, observed 2026-10-13T19:47Z)

Aggregate rank

#661 of 743 (score 43.8)

Relative to other models, GPT-5 nano (high) is strongest on Chess Puzzles (#29 of 122) and weakest on FrontierMath T4 (#48 of 50).

Across the 13 benchmarks where at least ten models are scored, GPT-5 nano (high) is in the top 10% on 0 and in the top half on 2.

Its cheapest listed input price, $0.050 per million tokens, is lower than that of 91% of the 427 models with a listed price.

GPT-5 nano (high) benchmark results

About GPT-5 nano (high)

GPT-5 nano (high) is a smaller, lower-cost model from OpenAI. It was released on 2025-08-07, according to the developer's announcement. The cheapest API listing we track charges $0.050 per million input tokens and $0.400 per million output tokens, through azure. We hold 13 benchmark results for it across 4 areas, and none is in the top tenth of its board.

Where GPT-5 nano (high) stands

On our aggregate score, GPT-5 nano (high) is #661 of 743 ranked models, in the lower half. Its score is 43.8; the scale is centred on 50, so it sits 6.2 points below the typical model. That beats 11% of the ranked field. Among the 88 ranked OpenAI models it is #78. The next one up is o3-mini (high) at 44.0. The next one down is GPT-4 (Mar 2023) at 43.7.

GPT-5 nano (high), benchmark by benchmark

Chess Puzzles puts GPT-5 nano (high) at 27.0%, #29 of 122. That beats 76% of the models tested, but 45.0 points short of GPT-6 Astra (max) (72.0%). The results either side are grok-4-0709 (28.0%) and Kimi K2.6 (26.0%). It covers chess tactics puzzles. The board median is 12.5%, so GPT-5 nano (high) is 14.5 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

On OTIS Mock AIME, GPT-5 nano (high) reaches 81.1%. It ranks #65 of 189, ahead of 66% of the field, and 18.9 points behind Claude Fable 5.1 (max) at 100.0%. The results either side are o4-mini (high) (81.7%) and Claude Sonnet 5 (max) (80.0%). It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so GPT-5 nano (high) is 23.3 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

GPT-5 nano (high) lands below the middle on GPQA Diamond: 69.4%, #145 of 275, with 52% of the field scoring higher. GPT-6 Astra (max) leads at 95.8%; the closest result above is qwen3-235b-a22b at 70.7%. It covers expert-level science questions. The board median is 71.7%, so GPT-5 nano (high) is 2.3 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

GPT-5 nano (high) lands below the middle on ALE-Bench: 719, #43 of 79, with 53% of the field scoring higher. GPT-6 Astra (max) leads at 2,951; the closest result above is MiMo-V2-Flash at 738. It covers AtCoder heuristic-optimisation contests (performance rating). The board median is 764, so GPT-5 nano (high) is 45 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

GPT-5 nano (high) lands below the middle on WeirdML: 38.1%, #68 of 109, with 61% of the field scoring higher. GPT-6 Astra (max) leads at 93.3%; the closest result above is Qwen3-235B-A22B-Instruct-2507 at 38.7%. It covers unusual machine-learning tasks. The board median is 41.2%, so GPT-5 nano (high) is 3.1 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

GPT-5 nano (high) lands below the middle on DTBench: 62.7%, #92 of 146, with 62% of the field scoring higher. Claude Opus 5.5 (max) leads at 98.9%; the closest result above is GPT-4 (Jun 2023) at 62.7%. It covers dynamic tool-use benchmark. The board median is 68.8%, so GPT-5 nano (high) is 6.1 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Mystery Games is a weak spot. GPT-5 nano (high) scores 8.0%, #48 of 59, behind 80% of the models measured. The leader, GPT-6 Astra (max), reaches 84.0%. It covers deductive reasoning puzzles. The board median is 23.0%, so GPT-5 nano (high) is 15.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

LMCA is a weak spot. GPT-5 nano (high) scores 7.9%, #100 of 108, behind 92% of the models measured. The leader, Claude Opus 5.5 (max), reaches 68.2%. It covers language-model capability assessment. The board median is 29.2%, so GPT-5 nano (high) is 21.3 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

SimpleQA Verified is a weak spot. GPT-5 nano (high) scores 11.7%, #66 of 69, behind 94% of the models measured. The leader, gemini-4-argon-high, reaches 76.0%. It covers short factual questions. The board median is 41.0%, so GPT-5 nano (high) is 29.3 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

FrontierMath T4 is a weak spot. GPT-5 nano (high) scores 2.4%, #48 of 50, behind 94% of the models measured. The leader, GPT-6.1 Sol (max), reaches 100.0%. It covers research-level maths, tier 4. The board median is 30.5%, so GPT-5 nano (high) is 28.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

The full table above also lists the other 3 results, including boards with fewer than ten models, where a rank says little.

Strengths and weaknesses

Reasoning, maths and science: uneven, ahead of 31% of the field on average across 6 benchmarks. Its best board here is OTIS Mock AIME (#65 of 189) and its weakest is FrontierMath T4 (#48 of 50). General knowledge and chat: weak, ahead of 29% of the field on average across 5 benchmarks. Its best board here is Chess Puzzles (#29 of 122) and its weakest is SimpleQA Verified (#66 of 69).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so GPT-5 nano (high) costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on qwen2.5-32b-instruct ($0.060) and deepseek-r1-0528-qwen3-8b ($0.060).

Release history

The OpenAI model released before it was GPT-5 mini (high) on 2025-08-07, which scores 46.9. The next one was GPT-5-codex (high) on 2025-09-15, at 50.5.

Reading these numbers

The 13 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (12), ARC Prize official leaderboard (1). 1 of them come from an official leaderboard and 12 from datasets or third-party runs. The leaderboard also lists 5 other reasoning-effort settings of this model, which redirect here. This page shows the highest.

GPT-5 nano (high) is also listed at 5 other reasoning-effort settings; this page shows the highest-effort row, which is the one the leaderboard keeps by default.

Models ranked nearby

• #658 ollama/hermes3 (43.9)
• #659 Qwen2-7B (43.9)
• #660 gemini-1.5-pro-002 (43.9)
• #662 gemma-2-27b-it (43.8)
• #663 Magistral Small 1.2 (43.8)
• #664 GPT-4 (Mar 2023) (43.7)

More from OpenAI

• GPT-6 Astra (max) (#2, 64.0)
• GPT-6.1 Sol (max) (#5, 60.6)
• GPT-5.6 Sol (max) (#6, 60.1)
• GPT-6 Sol (max) (#7, 60.0)
• GPT-5.5 (xhigh) (#11, 58.0)
• GPT-5.6 Terra (max) (#12, 57.8)

Compare it on the live leaderboard → · How the aggregate score works