Sponsored byFusecodeEnterprise coding API for Claude Code, Codex, and model workflows.

Free
Models
Providers
Docs

Home
Models
Gpt Oss 20b

LMSpeed

The best API speed test tool

GitHub Twitter Email

Product

Features
Pricing
FAQ

Leaderboard

Overview
Speed Ranking
Latency Ranking
Health Ranking
Model Pricing
Model Speed
Reasoning
Coding

Models

All Models
GPT
Claude
Gemini
DeepSeek
Llama
Qwen

Free Models

All Free Models
Free GPT
Free Claude
Free Gemini
Free DeepSeek
Free Llama
Free Qwen

Tools

Speed Test
Provider Audit

Resources

Provider Directory
Documentation
GEO Checker
Viora
Botab
VidBee

Legal

Cookie Policy
Privacy Policy
Terms of Service

© 2026 LMSpeed All Rights Reserved.Made by Nexmoe with ❤️

gpt-oss-20b API Pricing & Benchmarks | LMSpeed

OpenAI

gpt-oss-20b API Benchmarks, Pricing & Provider Data

gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimiz...

Quality

#87of 96

19.0

LMSpeed score

Category Performance

Scores across benchmark categories on a 0-100 scale.

Coverage: 6 / 8
Avg. score: 52.0

#1Multimodal

100

#2Instruction following

65.1

#3Knowledge

40.6

#4Coding

39.3

#5

Specifications

Input and output token limits for this model, plus how it ranks on long-context understanding.

INPUT

131.1Ktokens

≈ 157.3 pages of text

8K128K1M4M

131.1K

Features

Technical Details

Input

Rankings

Excels at

Falls behind in

Detailed scores

Updated: Jul 9, 2026

Overall

1 metric

Overall score19.0#87 / 96

Overall score

LMSpeed rank: #87 / 96Updated: Jul 9, 2026Confidence: 1

8424.0

8520.0

8520.0

8719.0

87gpt-oss-20b19.0

8918.0

9014.0

9111.0

9111.0

9310.0

Speed & latency

2 metrics

Output speed225.5 tok/s#8 / 113Time to first token0.50 s#6 / 113

Output speed

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #8 / 113Updated: Jul 9, 2026Confidence: 4

4263.2 tok/s

Pricing

2 metrics

Input price$0.060/M#4 / 138Output price$0.200/M#4 / 138

Input price

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #4 / 138Updated: Jul 9, 2026Confidence: 4

1$0.030/M

Reasoning & knowledge

14 metrics

MMLU-Pro71.8%#91 / 113GPQA61.1%#110 / 152

MMLU-Pro

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #91 / 113Updated: Jul 9, 2026Confidence: 4

8774.2%

Coding & agents

13 metrics

LiveCodeBench65.2%#39 / 108SciCode34.0%#92 / 149

LiveCodeBench

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #39 / 108Updated: Jul 9, 2026Confidence: 4

3468.4%

Language & media

3 metrics

Multilingual score5.6#53 / 59Multimodal Grounded score8.8#70 / 70

Multilingual score

LMSpeed rank: #53 / 59Updated: Jul 9, 2026Confidence: 1

5011.2

51

Related model comparisons

GGPT-5.4 vs gpt-oss-20b

3 verified comparison points

GGPT-5 vs gpt-oss-20b

3 verified comparison points

CClaude Opus 4.6 vs gpt-oss-20b

3 verified comparison points

Data as of Jul 9, 2026, 08:15 AMRankings are based on community-submitted tests and periodic health probes. Advisory only, not official data.Standard benchmark data may include BenchLM and other public sources.

Build with LMSpeed Data

Free

Free public API for LLM pricing, benchmarks & provider data

Real-time pricing & availability

Speed & latency benchmarks

300+ models, 600+ providers

1,000 requests/day

View API Documentation

Reasoning

34

#6Agents

33.3

No data:MathMultilingual

Output

Total parameters

20B

Knowledge cutoff

2024-06-30

Documentation

Tokenizer

GPT

Architecture

text->text

Moderated

No

Supported parameters

frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p

5Qwen3.5 Omni Flash243.4 tok/s

6Gemini 2.5 Flash230.1 tok/s

7O3 Mini228.5 tok/s

8gpt-oss-20b225.5 tok/s

9o3 Mini High219.2 tok/s

10GPT-5.1 Codex Mini212.6 tok/s

11Gemini 3 Flash Preview208.1 tok/s

12Gemini 3 Flash206.9 tok/s

13Hy3 preview203.9 tok/s

Time to first token

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #6 / 113Updated: Jul 9, 2026Confidence: 4

2Phi 4 Multimodal Instruct0.34 s

3Ministral 30.46 s

4Granite 4.1 8B0.46 s

5R1 Distill Llama 70B0.49 s

6gpt-oss-20b0.50 s

70.50 s

80.50 s

90.51 s

100.51 s

110.52 s

2

$0.050/M

2Granite 4.1 8B$0.050/M

4gpt-oss-20b$0.060/M

5GLM-4.7 Flash$0.070/M

6Nemotron 3 Nano Omni$0.075/M

7Devstral Small$0.100/M

7Gemini 2.5 Flash Lite$0.100/M

7Gemini 2.5 Flash Lite Preview 09-2025$0.100/M

7GPT-4.1 Nano$0.100/M

Output price

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #4 / 138Updated: Jul 9, 2026Confidence: 4

1Granite 4.1 8B$0.100/M

2Qwen3.5$0.150/M

2Qwen3.5-9B$0.150/M

4gpt-oss-20b$0.200/M

4$0.200/M

6$0.280/M

6$0.280/M

8$0.300/M

8$0.300/M

8$0.300/M

88

GPT-4o (2024-05-13)

74.0%

89DeepSeek R1 Distill Qwen73.9%

90Gemini 2.0 Flash Lite 00172.4%

91gpt-oss-20b71.8%

92Phi 471.4%

93Grok-270.9%

94Qwen3 Coder 30B A3B Instruct70.6%

95Claude 3 Opus69.6%

96GPT-4 Turbo69.4%

GPQA

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #110 / 152Updated: Jul 9, 2026Confidence: 4

106Gemini 2.5 Flash Lite62.5%

107Gemini 2.0 Pro62.2%

108Qwen3 30B A3B61.6%

109DeepSeek R1 Distill Qwen61.5%

110gpt-oss-20b61.1%

11161.0%

11260.4%

11360.3%

11459.9%

11559.4%

HLE

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #105 / 150Updated: Jul 9, 2026Confidence: 4

101DeepSeek R1 Distill Qwen5.5%

102GPT-55.4%

103Nemotron 3 Nano Omni5.3%

104gpt-oss-120b5.2%

105gpt-oss-20b5.1%

1055.1%

1074.9%

1074.9%

1094.7%

1104.6%

Reasoning score

LMSpeed rank: #64 / 65Updated: Jul 9, 2026Confidence: 1

56GPT-5.4 Mini28.3

57DeepSeek V3.126.7

58Qwen325.7

59GPT-4o Mini23.4

60DeepSeek V321.6

6121.3

6219.8

6315.2

64gpt-oss-20b12.4

659.9

AA-LCR

LMSpeed rank: #60 / 67Updated: Jul 9, 2026Confidence: 1

56GPT-4.1 Mini42.3

57DeepSeek V3.239.0

58Mistral Large 334.7

59MiMo-V2-Flash31.3

60gpt-oss-20b30.7

6129.0

6226.3

6325.8

6424.3

6521.0

CritPt

LMSpeed rank: #33 / 54Updated: Jul 9, 2026Confidence: 1

33DeepSeek R11.4

33Gemini 2.5 Flash1.4

33Gemini 3 Flash1.4

33gpt-oss-20b1.4

371.1

371.1

371.1

371.1

371.1

Knowledge score

LMSpeed rank: #79 / 79Updated: Jul 9, 2026Confidence: 1

70Qwen331.6

71Claude 3 Opus30.5

72Grok 329.8

73Gemini 2.5 Flash27.1

74Claude 3 Haiku24.7

7518.6

7617.7

7716.0

7813.3

79gpt-oss-20b10.5

Artificial Analysis Intelligence Index

LMSpeed rank: #63 / 77Updated: Jul 9, 2026Confidence: 1

60O1 Pro18.9

61GLM-4.5 Air16.5

62Mistral Large 315.9

63gpt-oss-20b14.9

6414.8

6514.2

6614.1

6711.8

6811.2

AA-GPQA Diamond

LMSpeed rank: #56 / 74Updated: Jul 9, 2026Confidence: 1

52DeepSeek V3.275.1

54O174.7

55GLM-4.5 Air73.3

56gpt-oss-20b68.8

5768.3

5768.3

5968.0

6066.6

6166.4

AA-HLE

LMSpeed rank: #51 / 74Updated: Jul 9, 2026Confidence: 1

47DeepSeek V3.113.0

48Claude Opus 4.512.9

49Qwen3 Max11.1

50DeepSeek V3.210.5

51gpt-oss-20b9.8

528.7

538.0

547.7

557.0

566.8

AA-Omniscience Accuracy

LMSpeed rank: #63 / 68Updated: Jul 9, 2026Confidence: 1

59Qwen3.6 35B A3B18.9

60MiMo-V2-Omni18.7

61GPT-4.1 Mini17.5

62Claude 3 Haiku17.2

6315.5

63gpt-oss-20b15.5

6515.2

6614.6

6713.3

6813.2

AA-Omniscience Hallucination Rate

LMSpeed rank: #2 / 68Updated: Jul 9, 2026Confidence: 1

1DeepSeek V4 Flash95.8

2gpt-oss-20b94.1

3DeepSeek V4 Pro94.0

4DeepSeek V3.293.5

593.3

692.3

791.2

890.9

990.3

1090.2

AA-IFBench

LMSpeed rank: #36 / 70Updated: Jul 9, 2026Confidence: 1

32gpt-oss-120b69.0

33MiMo-V2-Pro68.8

35Qwen3.6 27B67.6

36gpt-oss-20b65.1

3764.4

3863.1

3962.2

4061.1

4158.6

Instruction Following score

LMSpeed rank: #66 / 66Updated: Jul 9, 2026Confidence: 1

57gpt-oss-120b45.9

58Claude 3 Haiku45.3

59Claude Opus 4.542.4

60Hy3 preview39.9

6132.1

6230.8

6327.2

6424.1

6523.3

66gpt-oss-20b20.3

36

67.9%

37Olmo 3 32B Think67.2%

38GPT-5.266.9%

39gpt-oss-20b65.2%

40Qwen3 Max65.1%

41Gemini 2.5 Flash Lite Preview 09-202564.1%

42Ring Flash 2.062.8%

43Qwen3 235B A22B62.2%

44Kimi K2 090561.0%

SciCode

Source: Artificial Analysis / gpt-oss-20b (low)LMSpeed rank: #92 / 149Updated: Jul 9, 2026Confidence: 4

89GLM-4.534.8%

90Kimi K234.5%

91Claude Haiku 4.534.4%

92Gemini 2.0 Flash34.0%

92gpt-oss-20b34.0%

9433.8%

9533.7%

9633.3%

9733.1%

9733.1%

Coding score

LMSpeed rank: #81 / 83Updated: Jul 9, 2026Confidence: 1

74Gemini 2.5 Flash20.4

75DeepSeek V3.120.1

76Gemini 1.5 Pro19.8

77GPT-4 Turbo18.0

78Claude 3 Opus17.4

7917.2

8016.3

81gpt-oss-20b14.5

828.7

836.5

React Native Evals

LMSpeed rank: #11 / 11Updated: Jul 9, 2026Confidence: 1

2Claude Opus 4.684.1

3Claude Opus 4.782.8

4Claude Sonnet 4.680.6

5Gemini 3.1 Pro78.9

774.8

871.6

971.5

1071.4

11gpt-oss-20b71.0

AA Coding Index

LMSpeed rank: #30 / 36Updated: Jul 9, 2026Confidence: 1

26gpt-oss-120b30.4

27Gemini 1.5 Pro23.6

28DeepSeek V323.0

29GPT-4 Turbo21.5

30gpt-oss-20b20.7

3120.2

3220.1

3319.5

3411.4

3511.1

Terminal-Bench Hard

LMSpeed rank: #60 / 69Updated: Jul 9, 2026Confidence: 1

54Mistral Large 315.9

58O112.9

59Gemini 2.5 Flash12.1

60gpt-oss-20b10.6

618.3

627.6

636.8

636.8

636.8

AA-SciCode

LMSpeed rank: #60 / 74Updated: Jul 9, 2026Confidence: 1

56O135.8

56Qwen3.6 35B A3B35.8

58DeepSeek V335.4

60gpt-oss-20b34.4

6133.3

6233.1

6331.9

6430.6

6529.9

Agentic score

LMSpeed rank: #75 / 81Updated: Jul 9, 2026Confidence: 1

72GPT-4.1 Nano21.4

73Grok 321.0

74Claude 3 Haiku20.2

75gpt-oss-20b17.6

7716.9

7816.3

7916.0

8015.7

8113.8

AA Agentic Index

LMSpeed rank: #27 / 32Updated: Jul 9, 2026Confidence: 1

23MiMo-V2-Flash12.0

24Gemini 2.5 Pro7.1

25Gemini 3.1 Flash Lite6.2

26Mistral Large 35.5

27gpt-oss-20b3.1

281.7

291.6

301.2

311.1

321.0

APEX-Agents-AA

LMSpeed rank: #16 / 16Updated: Jul 9, 2026Confidence: 1

7GPT-5.4 Nano24.9

8DeepSeek V4 Pro24.3

10GLM-514.5

11Gemini 3.1 Flash Lite12.2

1211.5

1310.6

143.1

152.4

16gpt-oss-20b0.7

Tau2-Telecom

LMSpeed rank: #50 / 68Updated: Jul 9, 2026Confidence: 1

46Claude Opus 4.774.0

47gpt-oss-120b65.8

48O162.6

50gpt-oss-20b60.2

5154.1

5252.9

5352.3

5447.1

5546.5

GDPval-AA

LMSpeed rank: #27 / 27Updated: Jul 9, 2026Confidence: 1

19Qwen3.5-122B-A10B24.0

20Gemini 3.1 Pro23.1

2216.3

2314.8

248.0

256.9

266.4

27gpt-oss-20b2.7

GDPval-AA

LMSpeed rank: #28 / 33Updated: Jul 9, 2026Confidence: 1

24gpt-oss-120b796.0

25Gemini 2.5 Pro661.0

26Gemini 3.1 Flash Lite637.0

27Mistral Large 3629.0

28gpt-oss-20b554.0

29496.0

30216.0

31212.0

3279.0

3332.0

8.4

51Phi 48.4

53GPT-4.1 Nano5.6

53gpt-oss-20b5.6

55Llama 4 Scout2.8

57GLM-4.5 Air1.0

Multimodal Grounded score

LMSpeed rank: #70 / 70Updated: Jul 9, 2026Confidence: 1

61O1 Pro24.4

62gpt-oss-120b24.1

62Phi 424.1

64DeepSeek R122.4

65Grok 315.7

6613.3

6713.0

6812.1

6910.1

70gpt-oss-20b8.8

Design Arena Website

LMSpeed rank: #48 / 50Updated: Jul 9, 2026Confidence: 1

41Gemini 2.5 Flash1158.0

42Kimi K21093.0

43GPT-4.11081.0

44O31079.0

45GPT-4.1 Mini1040.0

461015.0

471011.0

48gpt-oss-20b895.0

49874.0

50793.0

Claude Opus 4.7

Nemotron 3 Super

Gemini 2.5 Flash Lite

Qwen3 Next 80B A3B Instruct

Hermes 3 70B Instruct

DeepSeek V4 Flash

Nemotron 3 Nano Omni

Olmo 3 32B Think

Claude 3.5 Sonnet

Gemini 2.0 Flash

Gemini 2.5 Flash Lite Preview 09-2025

Claude Sonnet 4

Gemini 3.1 Flash Lite

Gemini 2.5 Flash

Claude Sonnet 4

Gemini 2.5 Flash

Mistral Large 3

Gemini 2.5 Flash

Qwen3.6 35B A3B

Claude Opus 4.8

Claude Opus 4.7 Max

Mistral Medium 3.1

GPT-4o (2024-08-06)

Mistral Large 3

Claude Sonnet 4

Gemini 3.1 Flash Lite

Mistral Large 3