Pairwise benchmark snapshot · Aug 17, 2026

Llama 3.1 405B vs Yi-Lightning benchmark

Name: Llama 3.1 405B vs Yi-Lightning AI benchmark comparison
Creator: CompareLLM
License: https://creativecommons.org/licenses/by/4.0/

In this head-to-head showdown, Yi-Lightning delivers higher overall intelligence and human-preferred responses (Elo 1,475 vs 1,370), while Yi-Lightning leads in SWE-bench software engineering benchmarks (63.4%), while Yi-Lightning is more budget-friendly at $0.14/1M per 1M output tokens. Review the complete breakdown below to determine which model best fits your performance and budget requirements.

Share Analysis:Share:

WhatsApp Telegram X LinkedIn

Executive Verdict Summary

Yi-Lightning holds empirical advantage (9 of 10 metrics)

Coding / SWESWE-bench

Yi-LightningTop code solve rate

ReasoningGPQA / Live

Yi-LightningTop logical accuracy

Best Value$/1M Tok

Yi-LightningLowest API billing cost

Speed / TokTok/Sec

Yi-LightningFastest stream rate

Verified Advantage Breakdown

10 benchmarks evaluated across capability, speed, and cost

Llama 3.1 405B (1)Yi-Lightning (9)

Llama 3.1 405B

1 of 10 Wins

✓Context window

Yi-Lightning

9 of 10 Wins

✓Preference Elo✓Coding Elo✓LiveBench✓SWE-bench✓GPQA Diamond✓Time to first token✓Output speed✓Input price✓Output price

Yi-Lightning

Δ 105 Arena Elo pts

Throughput LeaderYi-Lightning Wins

Yi-Lightning

130 tok/s

Value per Dollar LeaderYi-Lightning Wins

Yi-Lightning

$0.14 / 1M output

Capability Percentiles

Relative percentile scores computed across all active models in the benchmark catalog.

Multi-Dimensional Capability Radar

Leaders Matchup Capability Radar

Comparing top standard models across 6 key capability dimensions. Tap any spoke or dot to inspect.

Tap any node to inspect

0–100 %ile

Overall Matchup Breakdown

Category wins across reasoning intelligence, generation speed, and token cost.

Llama 3.1 405B (1)vsYi-Lightning (9)

Llama 3.1 405B: 1W (10%)Yi-Lightning: 9W (90%)

← Llama 3.1 405BYi-Lightning →

🏆Yi-Lightning Leads(5/5)

Intelligence & Reasoning

Preference Elo, Coding proficiency, SWE-bench & LiveBench accuracy

Benchmark

Llama 3.1 405BvsYi-Lightning

Preference Elo

1,370vs1,475+105

Coding Elo

1,385vs1,480+95

SWE-bench

58.4%vs63.4%+5%

LiveBench

61.2%vs65%+3.8%

GPQA Diamond

75.2%vs78%+2.8%

Llama 3.1 405B: 0WYi-Lightning: 5W

🏆Yi-Lightning Leads(2/2)

Speed & Latency

Generation throughput and time to first token responsiveness

Benchmark

Llama 3.1 405BvsYi-Lightning

Output speed

42 tok/svs130 tok/s+88 tok/s

Time to first token

360 msvs150 ms+210 ms

Llama 3.1 405B: 0WYi-Lightning: 2W

🏆Yi-Lightning Leads(2/3)

Pricing & Capacity

Cost per million tokens and max context window length

Benchmark

Llama 3.1 405BvsYi-Lightning

Output price

$3.5/1Mvs$0.14/1M+$3.36/1M

Input price

$1.25/1Mvs$0.14/1M+$1.11/1M

Context window

128kvs16k+112k

Llama 3.1 405B: 1WYi-Lightning: 2W

Side-by-Side Benchmark Matrix

Llama 3.1 405BYi-Lightning

Preference EloYi-Lightning +105

Llama 3.1 405B1,370

Yi-Lightning1,475

Coding EloYi-Lightning +95

Llama 3.1 405B1,385

Yi-Lightning1,480

LiveBenchYi-Lightning +3.8%

Llama 3.1 405B61.2%

Yi-Lightning65%

SWE-benchYi-Lightning +5%

Llama 3.1 405B58.4%

Yi-Lightning63.4%

GPQA DiamondYi-Lightning +2.8%

Llama 3.1 405B75.2%

Yi-Lightning78%

Time to first tokenYi-Lightning +210 ms

Llama 3.1 405B360 ms

Yi-Lightning150 ms

Output speedYi-Lightning +88 tok/s

Llama 3.1 405B42 tok/s

Yi-Lightning130 tok/s

Input priceYi-Lightning +$1.11/1M

Llama 3.1 405B$1.25/1M

Yi-Lightning$0.14/1M

Output priceYi-Lightning +$3.36/1M

Llama 3.1 405B$3.5/1M

Yi-Lightning$0.14/1M

Context windowLlama 3.1 405B +112k

Llama 3.1 405B128k

Yi-Lightning16k

Benchmark	Llama 3.1 405B	Yi-Lightning	Advantage Delta
Preference Elo	1,370 lmarena · Aug 17, 2026	1,475 lmarena · Aug 17, 2026	+105
Coding Elo	1,385 lmarena · Aug 17, 2026	1,480 lmarena · Aug 17, 2026	+95
LiveBench	61.2% livebench · Aug 17, 2026	65% livebench · Aug 17, 2026	+3.8%
SWE-bench	58.4% swebench · Aug 17, 2026	63.4% swebench · Aug 17, 2026	+5%
GPQA Diamond	75.2% seed-bootstrap · Aug 1, 2026	78% seed-bootstrap · Aug 1, 2026	+2.8%
Time to first token	360 ms seed-bootstrap · Aug 1, 2026	150 ms seed-bootstrap · Aug 1, 2026	+210 ms
Output speed	42 tok/s seed-bootstrap · Aug 1, 2026	130 tok/s seed-bootstrap · Aug 1, 2026	+88 tok/s
Input price	$1.25/1M seed-bootstrap · Aug 1, 2026	$0.14/1M seed-bootstrap · Aug 1, 2026	+$1.11/1M
Output price	$3.5/1M seed-bootstrap · Aug 1, 2026	$0.14/1M seed-bootstrap · Aug 1, 2026	+$3.36/1M
Context window	128k seed-bootstrap · Aug 1, 2026	16k seed-bootstrap · Aug 1, 2026	+112k

Workload Cost & Savings Calculator

Simulate monthly production API costs in USD (US Dollar).

Save 93% with Yi-Lightning

Monthly Volume50M tokens/mo

Quick Presets:

Token Ratio (In vs Out)70% In / 30% Out

RAG / Search (10% out)Coding / Chat (50% out)

Llama 3.1 405B$96.25 / mo

In: $43.75Out: $52.5

Yi-Lightning$7.00 / mo

In: $4.9Out: $2.1

Estimated Cost Delta

Yi-Lightning is estimated to save $89.25/month ($1,071/year).

Based on 50M tokens (30% gen)

Key Selection Recommendations

1
Yi-Lightning:Pick Yi-Lightning when repo-level coding accuracy is the constraint.
2
Yi-Lightning:Pick Yi-Lightning when time-to-first-token matters more than peak Elo.
3
Yi-Lightning:Pick Yi-Lightning as the default general assistant.

Recommended Workload Routing

Repo / coding agentsYi-Lightning

Higher SWE-bench (63.4%).

High-volume chatYi-Lightning

Lower output list price ($0.14/1M).

Voice / low-latency UIYi-Lightning

Lower TTFT (150 ms).

Long-document RAGLlama 3.1 405B

Larger window (128k).

Screenshots / visionYi-Lightning

Yi-Lightning is the side marked multimodal in the catalog.

Target Workload	Recommended Pick	Evaluation Rationale
Repo / coding agents	Yi-Lightning	Higher SWE-bench (63.4%).
High-volume chat	Yi-Lightning	Lower output list price ($0.14/1M).
Voice / low-latency UI	Yi-Lightning	Lower TTFT (150 ms).
Long-document RAG	Llama 3.1 405B	Larger window (128k).
Screenshots / vision	Yi-Lightning	Yi-Lightning is the side marked multimodal in the catalog.

Cast Your Matchup Vote

0 total votes

Community Discussions (0)

No comments posted on this matchup yet. Be the first to share an evaluation note!

Social Share Card Preview

Dynamic OpenGraph banner generated at /compare/llama-3-1-405b-vs-yi-lightning/opengraph-image

Save PNG

CompareLLM AI Matrix

Verified Head-to-Head Telemetry

LIVE BENCHMARK

Llama 3.1 405B

Elo 1,370

SWE-bench58.4%

Speed (tok/s)42 tok/s

Input / 1M Tokens$1.25/1M

Output / 1M Tokens$3.5/1M

01.AI

Yi-Lightning

Elo 1,475

SWE-bench63.4%

Speed (tok/s)130 tok/s

Input / 1M Tokens$0.14/1M

Output / 1M Tokens$0.14/1M

Frequently Asked Questions

Which is better overall, Llama 3.1 405B or Yi-Lightning?: Yi-Lightning has the higher preference Elo in our latest snapshot (1,475). “Better” still depends on coding, price, and latency — see the table.
Which is better at coding, Llama 3.1 405B or Yi-Lightning?: Yi-Lightning leads SWE-bench at 63.4% vs 58.4%. SWE-bench is one harness, not your repo.
Which is cheaper to run in production?: Yi-Lightning output tokens are $0.14/1M versus $3.5/1M. Input prices and retry rates still move the real bill.
Which is faster, Llama 3.1 405B or Yi-Lightning?: Yi-Lightning has the lower time-to-first-token (150 ms vs 360 ms). Tokens/sec is a separate column if you care about long completions.
Which has the larger context window?: Llama 3.1 405B accepts 128k tokens versus 16k.
Can I self-host either model?: Llama 3.1 405B is marked open-weights in our catalog. The other side is a closed API. Check the provider license before you ship weights.
Are these scores from CompareLLM’s own evals?: No. V1 aggregates public leaderboards (and optional first-party latency pings). Each cell names its source and date. Read /methodology.
What does preference Elo mean on this page?: Preference Elo is a crowd vote from LMArena / Arena. People see two hidden answers and pick the one they like more. The model that wins more often gets a higher Elo. That means people preferred it — not that it passed a school test. It is not SWE-bench, not accuracy, and not a number we invent. Llama 3.1 405B and Yi-Lightning Elo cells are dated snapshots from the named source — we do not compute Arena ourselves.

Similar Head-to-Head Comparisons

All Llama 3.1 405B matchups →|All Yi-Lightning matchups →

Llama 3.1 405B vs Llama 3.1 70B (next llama)Yi-Lightning vs Yi-Large (previous other)Yi-Lightning vs Ernie 4.5 Turbo (next other)Llama 4 Maverick vs Yi-Lightning Llama 4 Scout vs Llama 3.1 405B Llama 3.3 70B vs Llama 3.1 405B

Llama 3.1 405B vs Yi-Lightning benchmark

Capability Percentiles

Relative percentile scores computed across all active models in the benchmark catalog.

Multi-Dimensional Capability Radar

Leaders Matchup Capability Radar

Comparing top standard models across 6 key capability dimensions. Tap any spoke or dot to inspect.

Tap any node to inspect

0–100 %ile

Side-by-Side Benchmark Matrix

Llama 3.1 405BYi-Lightning

Preference EloYi-Lightning +105

Llama 3.1 405B1,370

Yi-Lightning1,475

Coding EloYi-Lightning +95

Llama 3.1 405B1,385

Yi-Lightning1,480

LiveBenchYi-Lightning +3.8%

Llama 3.1 405B61.2%

Yi-Lightning65%

SWE-benchYi-Lightning +5%

Llama 3.1 405B58.4%

Yi-Lightning63.4%

GPQA DiamondYi-Lightning +2.8%

Llama 3.1 405B75.2%

Yi-Lightning78%

Time to first tokenYi-Lightning +210 ms

Llama 3.1 405B360 ms

Yi-Lightning150 ms

Output speedYi-Lightning +88 tok/s

Llama 3.1 405B42 tok/s

Yi-Lightning130 tok/s

Input priceYi-Lightning +$1.11/1M

Llama 3.1 405B$1.25/1M

Yi-Lightning$0.14/1M

Output priceYi-Lightning +$3.36/1M

Llama 3.1 405B$3.5/1M

Yi-Lightning$0.14/1M

Context windowLlama 3.1 405B +112k

Llama 3.1 405B128k

Yi-Lightning16k

Benchmark

Llama 3.1 405B

Yi-Lightning

Advantage Delta

Preference Elo

1,370

lmarena · Aug 17, 2026

1,475

lmarena · Aug 17, 2026

+105

Coding Elo

1,385

lmarena · Aug 17, 2026

1,480

lmarena · Aug 17, 2026

+95

LiveBench

61.2%

livebench · Aug 17, 2026

65%

livebench · Aug 17, 2026

+3.8%

SWE-bench

58.4%

swebench · Aug 17, 2026

63.4%

swebench · Aug 17, 2026

+5%

GPQA Diamond

75.2%

seed-bootstrap · Aug 1, 2026

78%

seed-bootstrap · Aug 1, 2026

+2.8%

Time to first token

360 ms

seed-bootstrap · Aug 1, 2026

150 ms

seed-bootstrap · Aug 1, 2026

+210 ms

Output speed

42 tok/s

seed-bootstrap · Aug 1, 2026

130 tok/s

seed-bootstrap · Aug 1, 2026

+88 tok/s

Input price

$1.25/1M

seed-bootstrap · Aug 1, 2026

$0.14/1M

seed-bootstrap · Aug 1, 2026

+$1.11/1M

Output price

$3.5/1M

seed-bootstrap · Aug 1, 2026

$0.14/1M

seed-bootstrap · Aug 1, 2026

+$3.36/1M

Context window

128k

seed-bootstrap · Aug 1, 2026

16k

seed-bootstrap · Aug 1, 2026

+112k

Recommended Workload Routing

Repo / coding agentsYi-Lightning

Higher SWE-bench (63.4%).

High-volume chatYi-Lightning

Lower output list price ($0.14/1M).

Voice / low-latency UIYi-Lightning

Lower TTFT (150 ms).

Long-document RAGLlama 3.1 405B

Larger window (128k).

Screenshots / visionYi-Lightning

Yi-Lightning is the side marked multimodal in the catalog.

Target Workload	Recommended Pick	Evaluation Rationale
Repo / coding agents	Yi-Lightning	Higher SWE-bench (63.4%).
High-volume chat	Yi-Lightning	Lower output list price ($0.14/1M).
Voice / low-latency UI	Yi-Lightning	Lower TTFT (150 ms).
Long-document RAG	Llama 3.1 405B	Larger window (128k).
Screenshots / vision	Yi-Lightning	Yi-Lightning is the side marked multimodal in the catalog.

Frequently Asked Questions

Which is better overall, Llama 3.1 405B or Yi-Lightning?

Yi-Lightning has the higher preference Elo in our latest snapshot (1,475). “Better” still depends on coding, price, and latency — see the table.

Which is better at coding, Llama 3.1 405B or Yi-Lightning?

Yi-Lightning leads SWE-bench at 63.4% vs 58.4%. SWE-bench is one harness, not your repo.

Which is cheaper to run in production?

Yi-Lightning output tokens are $0.14/1M versus $3.5/1M. Input prices and retry rates still move the real bill.

Which is faster, Llama 3.1 405B or Yi-Lightning?

Yi-Lightning has the lower time-to-first-token (150 ms vs 360 ms). Tokens/sec is a separate column if you care about long completions.

Which has the larger context window?

Llama 3.1 405B accepts 128k tokens versus 16k.

Can I self-host either model?

Llama 3.1 405B is marked open-weights in our catalog. The other side is a closed API. Check the provider license before you ship weights.

Are these scores from CompareLLM’s own evals?

No. V1 aggregates public leaderboards (and optional first-party latency pings). Each cell names its source and date. Read /methodology.

What does preference Elo mean on this page?

Preference Elo is a crowd vote from LMArena / Arena. People see two hidden answers and pick the one they like more. The model that wins more often gets a higher Elo. That means people preferred it — not that it passed a school test. It is not SWE-bench, not accuracy, and not a number we invent. Llama 3.1 405B and Yi-Lightning Elo cells are dated snapshots from the named source — we do not compute Arena ourselves.