ZMIME
Comparison · 3 models · Updated Nov 18, 2026

DeepSeek R1 0528 vs Llama-3.1-70B-Instruct vs WizardLM 2 8x22B

DeepSeek R1 0528 comes out ahead, 63 to 39 and 32 on our weighted score, though WizardLM 2 8x22B is 47% cheaper per token.

Price in rating: Excluded · Include price

  1. Our pick

    DeepSeek

    DeepSeek R1 0528

    Released May 28, 2025

    63/100
    • ECI140.8
    • Price$0.512 / $2.19
    • Context164K
  2. Meta

    Llama-3.1-70B-Instruct

    Released Jul 23, 2024

    39/100
    • ECI124.7
    • Price$0.72 / $0.72
    • Context128K
  3. Microsoft

    WizardLM 2 8x22B

    Released Apr 15, 2024

    32/100
    • ECI—
    • Price$0.493 / $0.493
    • Context66K
01 — Verdict

DeepSeek R1 0528 is our pick

DeepSeek R1 0528 is the better all-round choice, scoring 63/100 against Llama-3.1-70B-Instruct (39) and WizardLM 2 8x22B (32). It leads on capability, inputs & features and context window. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond), because WizardLM 2 8x22B has no Capabilities Index score yet.

  • CapabilityDeepSeek R1 0528Shared benchmarks: DeepSeek R1 0528 76.3% · Llama-3.1-70B-Instruct 44.2% · WizardLM 2 8x22B 43.4%
  • Lowest priceWizardLM 2 8x22BWizardLM 2 8x22B $0.493 · Llama-3.1-70B-Instruct $0.72 · DeepSeek R1 0528 $0.932 per 1M tokens (3:1 blend)
  • Longest contextDeepSeek R1 0528DeepSeek R1 0528 163,840 · Llama-3.1-70B-Instruct 128,000 · WizardLM 2 8x22B 65,536 tokens
  • Widest inputsSame inputsDeepSeek R1 0528: Text · Llama-3.1-70B-Instruct: Text · WizardLM 2 8x22B: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightDeepSeek R1 0528Llama-3.1-70B-InstructWizardLM 2 8x22B
CapabilityShared benchmarks67%764443
Inputs & features20%35250
Context window13%363222
Overall100%63/10039/10032/100
02 — Side by side

Every spec in one table

Highlighted cells lead their row. Dashes mean the data is not published.

DeepSeek R1 0528 vs Llama-3.1-70B-Instruct vs WizardLM 2 8x22B specifications side by side
SpecificationDeepSeek R1 0528DeepSeekLlama-3.1-70B-InstructMetaWizardLM 2 8x22BMicrosoft
Capability
Capabilities Index (ECI)140.8 (best)124.7—
ECI rank#113 of 158 (best)#146 of 158—
GPQA DiamondGraduate-level science questions76.3% (best)44.2%43.4%
OTIS Mock AIME 2024–2025Competition mathematics66.4% (best)3.6%—
Price per million tokens
Input$0.512$0.72$0.493 (best)
Output$2.19$0.72$0.493 (best)
Cached input———
Blended (3:1)$0.932$0.72$0.493 (best)
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 16 providersMedian of 7 providersMedian of 3 providers
Limits
Context window163,840 tokens (best)128,000 tokens65,536 tokens
Max output32,768 tokens (best)4,096 tokens8,000 tokens
Inputs and features
TextYesYesYes
ImagesNoNoNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningYesNoNo
Tool callingYesYesNo
Structured outputNoNoNo
Availability
WeightsOpenMITOpenOpenApache-2.0
API model ID———
API providers16 (best)73
ReleasedMay 28, 2025Jul 23, 2024Apr 15, 2024
Knowledge cutoff—Dec 2023—
03 — Cost

What would a month cost?

Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.

  • DeepSeek R1 0528$9.50
  • Llama-3.1-70B-Instruct$8.64
  • WizardLM 2 8x22B$5.92
04 — Questions

Which should you choose?

Which is better: DeepSeek R1 0528, Llama-3.1-70B-Instruct or WizardLM 2 8x22B?

DeepSeek R1 0528 is the better all-round choice, scoring 63/100 against Llama-3.1-70B-Instruct (39) and WizardLM 2 8x22B (32). It leads on capability, inputs & features and context window. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond), because WizardLM 2 8x22B has no Capabilities Index score yet.

Which is cheaper, DeepSeek R1 0528, Llama-3.1-70B-Instruct or WizardLM 2 8x22B?

WizardLM 2 8x22B is cheaper at $0.493 input / $0.493 output per million tokens (median across 3 API providers). Llama-3.1-70B-Instruct costs $0.72 input / $0.72 output per million tokens (median across 7 API providers); DeepSeek R1 0528 costs $0.512 input / $2.19 output per million tokens (median across 16 API providers). At a typical mix of three input tokens to one output token, that is $0.493 per million tokens for WizardLM 2 8x22B versus $0.72 for Llama-3.1-70B-Instruct (1.5× as much) and $0.932 for DeepSeek R1 0528 (1.9× as much).

Which scores higher on benchmarks?

Not every model here has a Capabilities Index score, so we compare the average of shared benchmarks (GPQA Diamond): DeepSeek R1 0528 76.3%, Llama-3.1-70B-Instruct 44.2% and WizardLM 2 8x22B 43.4%. On individual benchmarks: GPQA Diamond — DeepSeek R1 0528 76.3%, Llama-3.1-70B-Instruct 44.2%, WizardLM 2 8x22B 43.4%.

Which is better for coding?

There are no published SWE-bench Verified results for DeepSeek R1 0528, Llama-3.1-70B-Instruct and WizardLM 2 8x22B yet, so there is no like-for-like coding score. On overall capability, DeepSeek R1 0528 leads, which tends to carry over to coding, but test on your own codebase. Note that WizardLM 2 8x22B does not support tool calling, which most coding agents need.

Which has the bigger context window?

DeepSeek R1 0528 has the largest context window at 163,840 tokens, against 128,000 for Llama-3.1-70B-Instruct and 65,536 for WizardLM 2 8x22B. Maximum output per response: DeepSeek R1 0528 up to 32,768, Llama-3.1-70B-Instruct up to 4,096, WizardLM 2 8x22B up to 8,000 tokens.

Which can read images, PDFs, audio or video?

DeepSeek R1 0528 accepts text; Llama-3.1-70B-Instruct accepts text; WizardLM 2 8x22B accepts text. They handle the same number of input types.

Are any of these open source?

Yes, all three publish their weights (MIT and Apache-2.0), so you can self-host them.

Which is newer?

DeepSeek R1 0528 is the newest, released May 28, 2025. Llama-3.1-70B-Instruct came out Jul 23, 2024; WizardLM 2 8x22B came out Apr 15, 2024. Knowledge cutoff: Llama-3.1-70B-Instruct Dec 2023.

How do you decide the winner?

This rating uses capability (67%, independent benchmark results); inputs & features (20%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (13%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 18, 2026.