Llama 4 Maverick
Llama 4 Maverick, a 17 billion active parameter model with 128 experts, is the best multimodal model in its class, beating GPT-4o and Gemini 2.0 Flash across a broad range of widely reported benchmarks, while achieving comparable results to the new DeepSeek v3 on reasoning and coding—at less than half the active parameters. Llama 4 Maverick offers a best-in-class performance to cost ratio with an experimental chat version scoring ELO of 1417 on LMArena.
- Vision
- Structured Output
Added Sep 5, 2025
Model weightsPricing
Auto routing · per 1M tokens- Input
- $0.406
- Output
- $1.14
- Cache read
- $0.216
Specifications
- Context window
- 1M
- Max output
- 65.5K
- Parameters
- 400B / 17B
- Total / active
Benchmarks
Benchmarks
Sourced from LMArena.
Arena Score
1327.0
Overall Rank
#254 / 421
Votes
39,397
Confidence Interval
1322.7 - 1331.3
Category Scores
Coding
#246 / 416
6,934 votes
1372.5
Math
#236 / 404
2,820 votes
1316.9
Longer Query
#249 / 399
7,065 votes
1334.9
Creative Writing
#230 / 419
5,144 votes
1307.3
Instruction Following
#255 / 421
10,489 votes
1313.6
Hard Prompts
#249 / 421
15,851 votes
1338.5
Additional Categories22
German
#195 / 320
1,028 votes
1330.6
Japanese
#199 / 288
935 votes
1257.5
Polish
#201 / 244
2,675 votes
1330.3
Korean
#216 / 289
796 votes
1262.3
Spanish
#222 / 303
816 votes
1325.4
Russian
#234 / 385
3,099 votes
1328.0
Expert
#242 / 371
2,013 votes
1325.5
Industry Legal And Government
#242 / 393
2,581 votes
1341.2
Industry Medicine And Healthcare
#242 / 390
2,290 votes
1344.5
French
#244 / 301
596 votes
1298.4
Industry Entertainment And Sports And Media
#244 / 419
6,915 votes
1295.3
Industry Writing And Literature And Language
#244 / 420
8,719 votes
1312.0
Multi Turn
#244 / 419
6,677 votes
1326.4
Industry Mathematical
#245 / 401
2,562 votes
1319.1
Non English
#247 / 421
20,495 votes
1311.2
Industry Software And It Services
#249 / 421
12,275 votes
1358.0
Chinese
#251 / 395
2,594 votes
1334.5
Exclude Ties
#253 / 421
28,078 votes
1278.2
Hard Prompts English
#254 / 419
7,680 votes
1350.4
Industry Life And Physical And Social Science
#254 / 419
6,705 votes
1342.0
Industry Business And Management And Financial Operations
#257 / 414
6,120 votes
1317.3
English
#263 / 421
18,892 votes
1339.8
Published 2026-11-09 · Matched as Llama 4 Maverick 17B 128E Instruct · CC BY 4.0
LMArena DatasetLeaderboard results
| Benchmark | Score | Rank on NanoGPT |
|---|---|---|
Epoch AI Benchmarking Hub Epoch AI data, CC BY 4.0 · Updated Nov 13, 2026 | ||
Epoch Capabilities Index | 132.1 | #106 of 115 |
UGI Leaderboard UGI Leaderboard data, Apache-2.0 · Updated Oct 16, 2026 | ||
UGI Score | 41.1 | #72 of 149 |
UGI Willingness | 4.0 | #61 of 149 |
UGI Writing | 24.0 | #141 of 149 |
Filtered and ranked to models on NanoGPT.
Providers
Choose or restrict providers with the API. Provider-selection prices include the selection markup. Provider selection guide · Pricing API (JSON, USD per 1K tokens)
 | $0.406 | $1.14 | $0.216 | N/A | N/A |
| $0.406 | $1.14 | $0.216 | N/A | N/A |
| $0.367 | $1.21 | N/A | N/A | N/A | |
| $0.367 | $1.05 | $0.178 | N/A | N/A | |
| $0.556 | $1.70 | N/A | N/A | N/A |
Batch rows for this model use normal routing and cost the same as standard API requests. Set provider in a row to choose a provider.