Llama-3.1-405B-Instruct: benchmark scores, price and ranking
Llama-3.1-405B-Instruct from Meta AI ranks #668 of 743 LLMs on the LLMs Tiger aggregate benchmark score (43.6), based on 8 benchmark results captured up to 2026-10-13.
- Organisation
- Meta AI
- Type
- Open-weight / local model
- Open weights
- Yes
- Released
- 2024-07-23 (source: epoch)
- Cheapest API price
- $3.50 in / $3.50 out per million tokens (via together_ai, observed 2026-10-13T19:47Z)
- Aggregate rank
- #668 of 743 (score 43.6)
Relative to other models, Llama-3.1-405B-Instruct is strongest on Aider Code Editing (#24 of 92) and weakest on Cybench (#14 of 16).
Across the 8 benchmarks where at least ten models are scored, Llama-3.1-405B-Instruct is in the top 10% on 0 and in the top half on 1.
Its cheapest listed input price, $3.50 per million tokens, is lower than that of 11% of the 427 models with a listed price.
Llama-3.1-405B-Instruct benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| Aider Code Editing | 66.2% | #24 of 92 | Aider Aider Code Editing leaderboard | 2026-10-13 |
| TheAgentCompany | 7.4% | #8 of 13 | Epoch AI Benchmarking Hub | 2026-10-13 |
| DTBench | 61.4% | #97 of 146 | Epoch AI Benchmarking Hub | 2026-10-13 |
| GPQA Diamond | 50.9% | #184 of 275 | Epoch AI Benchmarking Hub | 2026-10-13 |
| OTIS Mock AIME | 9.7% | #138 of 189 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SimpleBench | 23.0% | #52 of 61 | Epoch AI Benchmarking Hub | 2026-10-13 |
| WeirdML | 21.4% | #94 of 109 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Cybench | 7.5% | #14 of 16 | Epoch AI Benchmarking Hub | 2026-10-13 |
About Llama-3.1-405B-Instruct
Llama-3.1-405B-Instruct is an open-weight model that can run on your own hardware from Meta AI. It was released on 2024-07-23, according to Epoch AI. The cheapest API listing we track charges $3.50 per million input tokens and $3.50 per million output tokens, through together_ai. We hold 8 benchmark results for it across 4 areas, and none is in the top tenth of its board.
Where Llama-3.1-405B-Instruct stands
On our aggregate score, Llama-3.1-405B-Instruct is #668 of 743 ranked models, in the lower half. Its score is 43.6; the scale is centred on 50, so it sits 6.4 points below the typical model. That beats 10% of the ranked field. Among the 28 ranked Meta AI models it is #18. The next one up is Llama-2-70b-chat at 45.3. The next one down is Llama-2-70b-chat-hf at 42.9.
Llama-3.1-405B-Instruct, benchmark by benchmark
Aider Code Editing puts Llama-3.1-405B-Instruct at 66.2%, #24 of 92. That beats 74% of the models tested, but 18.0 points short of Claude 3.5-Sonnet-20241022 (84.2%). The results either side are gpt-4-0314 (66.2%) and gemini-1.5-pro-002 (65.4%). It covers Aider official benchmark. The board median is 55.6%, so Llama-3.1-405B-Instruct is 10.6 points above it. Source: Aider Aider Code Editing leaderboard, captured 2026-10-13.
Llama-3.1-405B-Instruct lands below the middle on TheAgentCompany: 7.4%, #8 of 13, with 54% of the field scoring higher. gemini-2.5-flash-preview-09-2025 leads at 41.1%; the closest result above is GPT-4o (Nov 2024) at 8.6%. It covers work tasks in a simulated software company. The board median is 8.6%, so Llama-3.1-405B-Instruct is 1.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Llama-3.1-405B-Instruct lands below the middle on DTBench: 61.4%, #97 of 146, with 66% of the field scoring higher. Claude Opus 5.5 (max) leads at 98.9%; the closest result above is gpt-4-turbo at 61.6%. It covers dynamic tool-use benchmark. The board median is 68.8%, so Llama-3.1-405B-Instruct is 7.4 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Llama-3.1-405B-Instruct lands below the middle on GPQA Diamond: 50.9%, #184 of 275, with 67% of the field scoring higher. GPT-6 Astra (max) leads at 95.8%; the closest result above is mistral-large-2411 at 51.3%. It covers expert-level science questions. The board median is 71.7%, so Llama-3.1-405B-Instruct is 20.8 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Llama-3.1-405B-Instruct lands below the middle on OTIS Mock AIME: 9.7%, #138 of 189, with 72% of the field scoring higher. Claude Fable 5.1 (max) leads at 100.0%; the closest result above is grok-2-1212 at 11.5%. It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so Llama-3.1-405B-Instruct is 48.1 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
SimpleBench is a weak spot. Llama-3.1-405B-Instruct scores 23.0%, #52 of 61, behind 84% of the models measured. The leader, Claude Fable 5 (max), reaches 81.9%. It covers trick questions that test common sense. The board median is 41.4%, so Llama-3.1-405B-Instruct is 18.4 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
WeirdML is a weak spot. Llama-3.1-405B-Instruct scores 21.4%, #94 of 109, behind 85% of the models measured. The leader, GPT-6 Astra (max), reaches 93.3%. It covers unusual machine-learning tasks. The board median is 41.2%, so Llama-3.1-405B-Instruct is 19.8 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Cybench is a weak spot. Llama-3.1-405B-Instruct scores 7.5%, #14 of 16, behind 81% of the models measured. The leader, grok-4-0709, reaches 43.0%. It covers cybersecurity capture-the-flag tasks. The board median is 17.5%, so Llama-3.1-405B-Instruct is 10.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Reasoning, maths and science: weak, ahead of 25% of the field on average across 3 benchmarks. Its best board here is GPQA Diamond (#184 of 275) and its weakest is SimpleBench (#52 of 61). Agents and tool use: uneven, ahead of 36% of the field on average across 2 benchmarks. Its best board here is TheAgentCompany (#8 of 13) and its weakest is DTBench (#97 of 146). General knowledge and chat: weak, ahead of 13% of the field on average across 2 benchmarks. Its best board here is WeirdML (#94 of 109) and its weakest is Cybench (#14 of 16).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so Llama-3.1-405B-Instruct costs more 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 mistral-large-2402 ($8.00) and gpt-4-turbo-2024-04-09 ($10.00).
Release history
The Meta AI model released before it was Llama 3.1-405B on 2024-07-23, which scores 51.2. The next one was Llama 3.1-70B on 2024-07-23, at 47.5.
Reading these numbers
The 8 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (7), Aider Aider Code Editing leaderboard (1). 1 of them come from an official leaderboard and 7 from datasets or third-party runs.
Models ranked nearby
- #665 gemma-2-27b (43.7)
- #666 mistral-medium-2508 (43.7)
- #667 Claude 2 (43.7)
- #669 granite-3.1-2b-base (43.6)
- #670 deepseek-r1-0528-qwen3-8b (43.6)
- #671 ollama/granite3-dense:8b (43.6)
More from Meta AI
- Muse Spark 1.3 (max) (#18, 55.1)
- Muse Spark (#29, 53.7)
- Muse Spark 1.1 (xhigh) (#66, 52.4)
- Muse Spark 1.2 (xhigh) (#113, 51.5)
- Llama 3.1-405B (#142, 51.2)
- Muse Wonder (max) (#209, 50.1)
Compare it on the live leaderboard → · How the aggregate score works