All General Coding Reasoning Multimodal Safety Specialized Agentic Language Long Context Speed Text-to-Image Image Editing Text-to-Video Image-to-Video
Generation Tokens/sec incl. Reasoning (RightRank Probe)113 Inter-Token Latency Latency Output Tokens/sec163 Output Tokens/sec (RightRank Probe)165 TTFT163 TTFT (RightRank Probe)180 Time to First Answer Token (RightRank Probe)100 Tool Call TTFT (RightRank Probe)169 Tool Call Total Time (RightRank Probe)169 Total Time (RightRank Probe)181

Speed

Composite performance across 9 benchmarks · Normalized 0–100 · At least 5 scores required · Ranked by average performance · Top 50 of 180 eligible models

# Model Provider Generation ... Output Toke... Output Toke... TTFT TTFT (Right... Time to Fir... Tool Call T... Tool Call T... Total Time ... Avg Coverage Price Value
1 Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) Google 365.5 tok/s — 547.3 tok/s — 569 ms 569 ms — — 1.6 s
90.5
5/9 $1.75/M 51.7
2 Google: Gemini 2.5 Flash LiteRetiring Google 242.2 tok/s — 284.0 tok/s — 263 ms 263 ms 306 ms 313 ms 1.8 s
90.3
7/9 $0.50/M 180.7
3 Inception: Mercury 2 Inception 320.4 tok/s — 387.1 tok/s — 560 ms 560 ms 326 ms 337 ms 1.8 s
88.6
7/9 $1.00/M 88.6
4 Mistral: Ministral 3 3B 2512 Mistral AI 241.5 tok/s — 301.3 tok/s — 343 ms 343 ms 354 ms 362 ms 2.3 s
87.6
7/9 $0.20/M 438.1
5 Amazon: Nova Micro 1.0 Amazon 250.7 tok/s — 305.3 tok/s — 364 ms 364 ms 431 ms 570 ms 2.0 s
86.0
7/9 $0.18/M 491.7
6 Google: Gemini 3.1 Flash Lite Preview Google 210.6 tok/s — 245.5 tok/s — 384 ms 384 ms 433 ms 440 ms 2.7 s
84.0
7/9 $1.75/M 48.0
7 Google: Gemini 3.1 Flash Lite Google 203.3 tok/s — 246.6 tok/s — 386 ms 386 ms 449 ms 457 ms 2.8 s
83.6
7/9 $1.75/M 47.7
8 Inception: Mercury 2.5 Inception 418.7 tok/s 731.3 tok/s — 2.6 s — — 655 ms 701 ms 1.4 s
83.6
6/9 $0.19/M 440.2
9 Mistral: Voxtral Small 24B 2507 Mistral AI 147.4 tok/s — 165.7 tok/s — 292 ms 292 ms 513 ms 518 ms 2.8 s
81.3
7/9 $0.40/M 203.3
10 Amazon: Nova Lite 1.0 Amazon 173.5 tok/s — 193.1 tok/s — 351 ms 351 ms 346 ms 674 ms 3.4 s
81.1
7/9 $0.30/M 270.5
11 Mistral: Mistral Medium 3 Mistral AI 125.9 tok/s — 148.7 tok/s — 488 ms 488 ms 371 ms 372 ms 3.5 s
78.8
7/9 $2.40/M 32.8
12 Mistral: Mistral Medium 3.1 Mistral AI 124.8 tok/s — 151.0 tok/s — 518 ms 518 ms 403 ms 404 ms 2.9 s
78.7
7/9 $2.40/M 32.8
13 Mistral: Codestral 2508 Mistral AI 124.7 tok/s — 133.6 tok/s — 320 ms 320 ms 436 ms 440 ms 4.6 s
78.6
7/9 $1.20/M 65.5
14 Google: Gemini 2.5 FlashRetiring Google 155.6 tok/s — 197.1 tok/s — 665 ms 665 ms 432 ms 438 ms 3.4 s
78.4
7/9 $2.80/M 28.0
15 OpenAI: GPT-5.4 Nano OpenAI 161.1 tok/s — 195.6 tok/s — 481 ms 481 ms 572 ms 720 ms 3.1 s
78.2
7/9 $1.45/M 53.9
16 OpenAI: GPT-4o-mini OpenAI 136.0 tok/s — 153.9 tok/s — 533 ms 533 ms 376 ms 449 ms 4.4 s
77.4
7/9 $0.75/M 103.2
17 Amazon: Nova 2 Lite Amazon 170.1 tok/s — 209.9 tok/s — 735 ms 735 ms 510 ms 525 ms 3.5 s
77.4
7/9 $2.80/M 27.6
18 Mistral: Ministral 3 14B 2512 Mistral AI 107.8 tok/s — 116.3 tok/s — 314 ms 314 ms 427 ms 437 ms 5.1 s
77.3
7/9 $0.40/M 193.1
19 Qwen: Qwen3 VL 8B InstructHistorical Alibaba — — 152.6 tok/s — 420 ms — 538 ms 607 ms 4.3 s
77.0
5/9 $1.00/M 77.0
20 OpenAI: GPT-4.1 Mini OpenAI 157.0 tok/s — 178.4 tok/s — 486 ms 486 ms 612 ms 713 ms 3.8 s
76.8
7/9 $2.00/M 38.4
21 Google: Gemini 3.5 Flash Lite Google 182.6 tok/s 406.3 tok/s 242.2 tok/s 6.0 s 683 ms 683 ms 395 ms 402 ms 3.3 s
76.4
9/9 $2.80/M 27.3
22 Mistral: Devstral 2 2512 Mistral AI 119.7 tok/s — 147.6 tok/s — 521 ms 521 ms 529 ms 536 ms 3.4 s
76.3
7/9 $2.40/M 31.8
23 Cohere: Command A+ Cohere 102.2 tok/s — 98.2 tok/s — 180 ms 1.8 s 136 ms 463 ms 5.9 s
76.1
7/9 $1.80/M 42.3
24 OpenAI: GPT-3.5 Turbo OpenAI 111.5 tok/s — 127.5 tok/s — 422 ms 422 ms 575 ms 653 ms 3.2 s
75.8
7/9 $2.00/M 37.9
25 OpenAI: GPT-4o-mini (2024-07-18) OpenAI 128.4 tok/s — 146.8 tok/s — 511 ms 511 ms 530 ms 574 ms 4.2 s
75.7
7/9 $0.75/M 101.0
26 Qwen: Qwen3 VL 30B A3B Instruct Alibaba 112.2 tok/s — 131.3 tok/s — 481 ms 481 ms 573 ms 644 ms 3.0 s
75.6
7/9 $0.75/M 100.8
27 OpenAI: GPT-4.1 Nano OpenAI 155.8 tok/s — 194.6 tok/s — 596 ms 596 ms 764 ms 828 ms 3.4 s
75.5
7/9 $0.50/M 150.9
28 Qwen: Qwen3.5-Flash Alibaba 163.0 tok/s — 194.5 tok/s — 654 ms 654 ms 670 ms 1.1 s 3.7 s
74.3
7/9 $0.32/M 228.7
29 Mistral: Ministral 3 8B 2512 Mistral AI 81.2 tok/s — 84.9 tok/s — 279 ms 279 ms 429 ms 436 ms 7.4 s
74.1
7/9 $0.30/M 247.1
30 Qwen: Qwen3 30B A3B Instruct 2507 Alibaba 80.8 tok/s — 88.3 tok/s — 369 ms 369 ms 449 ms 501 ms 4.5 s
74.1
7/9 $0.40/M 185.2
31 Cohere: North Mini Code (free) Cohere — — 56.5 tok/s — 195 ms — 173 ms 689 ms 11.0 s
73.7
5/9 Free —
32 OpenAI: GPT-4o (2024-05-13) OpenAI — — 165.6 tok/s — 729 ms — 672 ms 744 ms 4.3 s
73.7
5/9 $20.00/M 3.7
33 OpenAI: GPT-3.5 Turbo 16k OpenAI 117.6 tok/s — 154.4 tok/s — 723 ms 723 ms 679 ms 753 ms 3.1 s
73.3
7/9 $7.00/M 10.5
34 Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview) Google 157.3 tok/s — 210.2 tok/s — 926 ms 926 ms — — 3.7 s
73.2
5/9 $3.50/M 20.9
35 Google: Gemini 3 Flash Preview Google 137.2 tok/s — 172.7 tok/s — 855 ms 855 ms 690 ms 698 ms 3.8 s
73.1
7/9 $3.50/M 20.9
36 Mistral: Mistral Medium 3.5 Mistral AI — 166.7 tok/s 127.4 tok/s 634 ms 548 ms — 664 ms 671 ms 3.3 s
73.1
7/9 $9.00/M 8.1
37 OpenAI: GPT-5.4 Mini OpenAI 117.4 tok/s — 135.4 tok/s — 639 ms 639 ms 548 ms 665 ms 4.7 s
72.9
7/9 $5.25/M 13.9
38 Perplexity: Sonar Perplexity 152.9 tok/s — 205.9 tok/s — 948 ms 948 ms — — 3.7 s
72.8
5/9 $2.00/M 36.4
39 Google: Gemini 3.8 Flash Google 177.0 tok/s — 364.3 tok/s — 1.4 s 3.2 s 721 ms 725 ms 3.4 s
72.7
7/9 $4.50/M 16.1
40 Google: Nano Banana 2 (Gemini 3.1 Flash Image) Google 156.0 tok/s — 212.4 tok/s — 1.0 s 1.0 s — — 3.7 s
72.5
5/9 $3.50/M 20.7
41 Qwen: Qwen3 Next 80B A3B InstructHistorical Alibaba 92.7 tok/s 163.7 tok/s 106.5 tok/s 573 ms 474 ms 474 ms 510 ms 539 ms 4.6 s
72.3
9/9 $1.20/M 60.3
42 Amazon: Nova Pro 1.0 Amazon 126.1 tok/s — 171.3 tok/s — 919 ms 919 ms 654 ms 882 ms 3.2 s
72.3
7/9 $4.00/M 18.1
43 Cohere: Command R7B (12-2024) Cohere 76.5 tok/s — 78.5 tok/s — 183 ms 183 ms — — 7.6 s
71.5
5/9 $0.19/M 381.2
44 OpenAI: GPT-6 Luna OpenAI 133.9 tok/s — 256.8 tok/s — 1.6 s 2.1 s 561 ms 708 ms 3.9 s
71.1
7/9 $0.60/M 118.5
45 SpaceXAI: Grok 4.20 SpaceXAI 89.0 tok/s — 99.6 tok/s — 699 ms 699 ms 428 ms 436 ms 6.7 s
70.9
7/9 $3.75/M 18.9
46 Qwen: Qwen3 VL 32B InstructHistorical Alibaba — — 92.7 tok/s — 512 ms — 713 ms 808 ms 4.7 s
70.9
5/9 $0.80/M 88.7
47 Qwen: Qwen3.6 Flash Alibaba 119.8 tok/s — 135.5 tok/s — 450 ms — 441 ms 1.6 s 5.0 s
70.4
6/9 $1.31/M 53.6
48 Google: Gemini 3.7 Flash Google 168.2 tok/s — 447.7 tok/s — 2.0 s 3.5 s 965 ms 972 ms 3.5 s
70.3
7/9 $4.50/M 15.6
49 OpenAI: GPT-4o OpenAI — — 158.3 tok/s — 895 ms — 914 ms 989 ms 4.5 s
70.1
5/9 $12.50/M 5.6
50 OpenAI: o3 Mini OpenAI 182.2 tok/s — 300.8 tok/s — 2.1 s 2.1 s 1.1 s 1.3 s 3.2 s
69.6
7/9 $5.50/M 12.6

Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) leads on raw performance, but Amazon: Nova Micro 1.0 delivers 851% more performance per dollar. That extra 4.5 points costs $0.35/M per point.

Access this data

JSON GET /api/v1 
[
  {
    "rank": 1,
    "model": "Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)",
    "provider": "Google",
    "avg_score": 90.5,
    "cost": 1.75,
    "cost_unit": "million_tokens",
    "value": 51.7
  },
  {
    "rank": 2,
    "model": "Google: Gemini 2.5 Flash Lite",
    "provider": "Google",
    "avg_score": 90.3,
    "cost": 0.5,
    "cost_unit": "million_tokens",
    "value": 180.7
  },
  {
    "rank": 3,
    "model": "Inception: Mercury 2",
    "provider": "Inception",
    "avg_score": 88.6,
    "cost": 1.0,
    "cost_unit": "million_tokens",
    "value": 88.6
  }
]
Curl 
$ curl -s https://rightrank.ai/api/v1/rankings/speed.json | jq '.[:3]'
CLI brew install rightrank 
$ rightrank top speed --limit 5
$ rightrank compare claude-opus-4.6 gpt-5.4-pro --category speed
$ rightrank value speed  # best performance per dollar
Claude Code Skill Skill 
# In Claude Code, use the RightRank skill:
/rightrank best model for speed
/rightrank compare top 5 by value
/rightrank cheapest model scoring above 80 on speed